[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
from typing import Annotated , Any , Literal
from fastmcp import Context
from mcp . types import ToolAnnotations
from services . registry import mcp_for_unity_tool
from services . tools import get_unity_instance_from_context
from transport . unity_transport import send_with_unity_instance
from transport . legacy . unity_connection import async_send_command_with_retry
# All possible actions grouped by component type
PARTICLE_ACTIONS = [
" particle_get_info " , " particle_set_main " , " particle_set_emission " , " particle_set_shape " ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
" particle_set_color_over_lifetime " , " particle_set_size_over_lifetime " ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
" particle_set_velocity_over_lifetime " , " particle_set_noise " , " particle_set_renderer " ,
" particle_enable_module " , " particle_play " , " particle_stop " , " particle_pause " ,
" particle_restart " , " particle_clear " , " particle_add_burst " , " particle_clear_bursts "
]
VFX_ACTIONS = [
# Asset management
" vfx_create_asset " , " vfx_assign_asset " , " vfx_list_templates " , " vfx_list_assets " ,
# Runtime control
" vfx_get_info " , " vfx_set_float " , " vfx_set_int " , " vfx_set_bool " ,
" vfx_set_vector2 " , " vfx_set_vector3 " , " vfx_set_vector4 " , " vfx_set_color " ,
" vfx_set_gradient " , " vfx_set_texture " , " vfx_set_mesh " , " vfx_set_curve " ,
" vfx_send_event " , " vfx_play " , " vfx_stop " , " vfx_pause " , " vfx_reinit " ,
" vfx_set_playback_speed " , " vfx_set_seed "
]
LINE_ACTIONS = [
" line_get_info " , " line_set_positions " , " line_add_position " , " line_set_position " ,
" line_set_width " , " line_set_color " , " line_set_material " , " line_set_properties " ,
" line_clear " , " line_create_line " , " line_create_circle " , " line_create_arc " , " line_create_bezier "
]
TRAIL_ACTIONS = [
" trail_get_info " , " trail_set_time " , " trail_set_width " , " trail_set_color " ,
" trail_set_material " , " trail_set_properties " , " trail_clear " , " trail_emit "
]
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
ALL_ACTIONS = [ " ping " ] + PARTICLE_ACTIONS + \
VFX_ACTIONS + LINE_ACTIONS + TRAIL_ACTIONS
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
@mcp_for_unity_tool (
description = """ Unified VFX management for Unity visual effects components.
Each action prefix requires a specific component on the target GameObject :
- ` particle_ * ` actions require * * ParticleSystem * * component
- ` vfx_ * ` actions require * * VisualEffect * * component ( + com . unity . visualeffectgraph package )
- ` line_ * ` actions require * * LineRenderer * * component
- ` trail_ * ` actions require * * TrailRenderer * * component
* * If the component doesn ' t exist, the action will FAIL
Before using this tool , either :
1. Use ` manage_gameobject ` with ` action = " get_components " ` to check if component exists
2. Use ` manage_gameobject ` with ` action = " add_component " , component_name = " ParticleSystem " ` ( or LineRenderer / TrailRenderer / VisualEffect ) to add the component first
3. Assign material to the component beforehand to avoid empty effects
* * TARGETING : * *
Use ` target ` parameter to specify the GameObject :
- By name : ` target = " Fire " ` ( finds first GameObject named " Fire " )
- By path : ` target = " Effects/Fire " ` with ` search_method = " by_path " `
- By instance ID : ` target = " 12345 " ` with ` search_method = " by_id " ` ( most reliable )
- By tag : ` target = " Player " ` with ` search_method = " by_tag " `
* * Component Types & Action Prefixes : * *
- ` particle_ * ` - ParticleSystem ( legacy particle effects )
- ` vfx_ * ` - Visual Effect Graph ( modern GPU particles , requires com . unity . visualeffectgraph )
- ` line_ * ` - LineRenderer ( lines , curves , shapes )
- ` trail_ * ` - TrailRenderer ( motion trails )
* * ParticleSystem Actions ( particle_ * ) : * *
- particle_get_info : Get particle system info
- particle_set_main : Set main module ( duration , looping , startLifetime , startSpeed , startSize , startColor , gravityModifier , maxParticles )
- particle_set_emission : Set emission ( rateOverTime , rateOverDistance )
- particle_set_shape : Set shape ( shapeType , radius , angle , arc , position , rotation , scale )
- particle_set_color_over_lifetime , particle_set_size_over_lifetime , particle_set_velocity_over_lifetime
- particle_set_noise : Set noise ( strength , frequency , scrollSpeed )
- particle_set_renderer : Set renderer ( renderMode , material )
- particle_enable_module : Enable / disable modules
- particle_play / stop / pause / restart / clear : Playback control
- particle_add_burst , particle_clear_bursts : Burst management
* * VFX Graph Actions ( vfx_ * ) : * *
- * * Asset Management : * *
- vfx_create_asset : Create a new VFX Graph asset file ( requires : assetName , optional : folderPath , template , overwrite )
- vfx_assign_asset : Assign a VFX asset to a VisualEffect component ( requires : target , assetPath )
- vfx_list_templates : List available VFX templates in project and packages
- vfx_list_assets : List all VFX assets in project ( optional : folder , search )
- * * Runtime Control : * *
- vfx_get_info : Get VFX info
- vfx_set_float / int / bool : Set exposed parameters
- vfx_set_vector2 / vector3 / vector4 : Set vector parameters
- vfx_set_color , vfx_set_gradient : Set color / gradient parameters
- vfx_set_texture , vfx_set_mesh : Set asset parameters
- vfx_set_curve : Set animation curve
- vfx_send_event : Send events with attributes ( position , velocity , color , size , lifetime )
- vfx_play / stop / pause / reinit : Playback control
- vfx_set_playback_speed , vfx_set_seed
* * LineRenderer Actions ( line_ * ) : * *
- line_get_info : Get line info
- line_set_positions : Set all positions
- line_add_position , line_set_position : Modify positions
- line_set_width : Set width ( uniform , start / end , curve )
- line_set_color : Set color ( uniform , gradient )
- line_set_material , line_set_properties
- line_clear : Clear positions
- line_create_line : Create simple line
- line_create_circle : Create circle
- line_create_arc : Create arc
- line_create_bezier : Create Bezier curve
* * TrailRenderer Actions ( trail_ * ) : * *
- trail_get_info : Get trail info
- trail_set_time : Set trail duration
- trail_set_width , trail_set_color , trail_set_material , trail_set_properties
- trail_clear : Clear trail
- trail_emit : Emit point ( Unity 2021.1 + ) """ ,
annotations = ToolAnnotations (
title = " Manage VFX " ,
destructiveHint = True ,
) ,
)
async def manage_vfx (
ctx : Context ,
action : Annotated [ str , " Action to perform. Use prefix: particle_, vfx_, line_, or trail_ " ] ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Target specification (common) - REQUIRED for most actions
# Using str | None to accept any string format
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
target : Annotated [ str | None ,
" Target GameObject with the VFX component. Use name (e.g. ' Fire ' ), path ( ' Effects/Fire ' ), instance ID, or tag. The GameObject MUST have the required component (ParticleSystem/VisualEffect/LineRenderer/TrailRenderer) for the action prefix. " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
search_method : Annotated [
Literal [ " by_id " , " by_name " , " by_path " , " by_tag " , " by_layer " ] | None ,
" How to find target: by_name (default), by_path (hierarchy path), by_id (instance ID - most reliable), by_tag, by_layer "
] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === PARTICLE SYSTEM PARAMETERS ===
# Main module - All use Any to accept string coercion from MCP clients
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
duration : Annotated [ Any ,
" [Particle] Duration in seconds (number or string) " ] = None ,
looping : Annotated [ Any ,
" [Particle] Whether to loop (bool or string ' true ' / ' false ' ) " ] = None ,
prewarm : Annotated [ Any ,
" [Particle] Prewarm the system (bool or string) " ] = None ,
start_delay : Annotated [ Any ,
" [Particle] Start delay (number or MinMaxCurve dict) " ] = None ,
start_lifetime : Annotated [ Any ,
" [Particle] Particle lifetime (number or MinMaxCurve dict) " ] = None ,
start_speed : Annotated [ Any ,
" [Particle] Initial speed (number or MinMaxCurve dict) " ] = None ,
start_size : Annotated [ Any ,
" [Particle] Initial size (number or MinMaxCurve dict) " ] = None ,
start_rotation : Annotated [ Any ,
" [Particle] Initial rotation (number or MinMaxCurve dict) " ] = None ,
start_color : Annotated [ Any ,
" [Particle/VFX] Start color [r,g,b,a] (array, dict, or JSON string) " ] = None ,
gravity_modifier : Annotated [ Any ,
" [Particle] Gravity multiplier (number or MinMaxCurve dict) " ] = None ,
simulation_space : Annotated [ Literal [ " Local " , " World " ,
" Custom " ] | None , " [Particle] Simulation space " ] = None ,
scaling_mode : Annotated [ Literal [ " Hierarchy " , " Local " ,
" Shape " ] | None , " [Particle] Scaling mode " ] = None ,
play_on_awake : Annotated [ Any ,
" [Particle] Play on awake (bool or string) " ] = None ,
max_particles : Annotated [ Any ,
" [Particle] Maximum particles (integer or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Emission
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
rate_over_time : Annotated [ Any ,
" [Particle] Emission rate over time (number or MinMaxCurve dict) " ] = None ,
rate_over_distance : Annotated [ Any ,
" [Particle] Emission rate over distance (number or MinMaxCurve dict) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Shape
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
shape_type : Annotated [ Literal [ " Sphere " , " Hemisphere " , " Cone " , " Box " ,
" Circle " , " Edge " , " Donut " ] | None , " [Particle] Shape type " ] = None ,
radius : Annotated [ Any ,
" [Particle/Line] Shape radius (number or string) " ] = None ,
radius_thickness : Annotated [ Any ,
" [Particle] Radius thickness 0-1 (number or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
angle : Annotated [ Any , " [Particle] Cone angle (number or string) " ] = None ,
arc : Annotated [ Any , " [Particle] Arc angle (number or string) " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Noise
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
strength : Annotated [ Any ,
" [Particle] Noise strength (number or MinMaxCurve dict) " ] = None ,
frequency : Annotated [ Any ,
" [Particle] Noise frequency (number or string) " ] = None ,
scroll_speed : Annotated [ Any ,
" [Particle] Noise scroll speed (number or MinMaxCurve dict) " ] = None ,
damping : Annotated [ Any ,
" [Particle] Noise damping (bool or string) " ] = None ,
octave_count : Annotated [ Any ,
" [Particle] Noise octaves 1-4 (integer or string) " ] = None ,
quality : Annotated [ Literal [ " Low " , " Medium " , " High " ]
| None , " [Particle] Noise quality " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Module control
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
module : Annotated [ str | None ,
" [Particle] Module name to enable/disable " ] = None ,
enabled : Annotated [ Any ,
" [Particle] Enable/disable module (bool or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Burst
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
time : Annotated [ Any ,
" [Particle/Trail] Burst time or trail duration (number or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
count : Annotated [ Any , " [Particle] Burst count (integer or string) " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
min_count : Annotated [ Any ,
" [Particle] Min burst count (integer or string) " ] = None ,
max_count : Annotated [ Any ,
" [Particle] Max burst count (integer or string) " ] = None ,
cycles : Annotated [ Any ,
" [Particle] Burst cycles (integer or string) " ] = None ,
interval : Annotated [ Any ,
" [Particle] Burst interval (number or string) " ] = None ,
probability : Annotated [ Any ,
" [Particle] Burst probability 0-1 (number or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Playback
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
with_children : Annotated [ Any ,
" [Particle] Apply to children (bool or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === VFX GRAPH PARAMETERS ===
# Asset management
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
asset_name : Annotated [ str | None ,
" [VFX] Name for new VFX asset (without .vfx extension) " ] = None ,
folder_path : Annotated [ str | None ,
" [VFX] Folder path for new asset (default: Assets/VFX) " ] = None ,
template : Annotated [ str | None ,
" [VFX] Template name for new asset (use vfx_list_templates to see available) " ] = None ,
asset_path : Annotated [ str | None ,
" [VFX] Path to VFX asset to assign (e.g. Assets/VFX/MyEffect.vfx) " ] = None ,
overwrite : Annotated [ Any ,
" [VFX] Overwrite existing asset (bool or string) " ] = None ,
folder : Annotated [ str | None ,
" [VFX] Folder to search for assets (for vfx_list_assets) " ] = None ,
search : Annotated [ str | None ,
" [VFX] Search pattern for assets (for vfx_list_assets) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Runtime parameters
parameter : Annotated [ str | None , " [VFX] Exposed parameter name " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
value : Annotated [ Any ,
" [VFX] Parameter value (number, bool, array, or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
texture_path : Annotated [ str | None , " [VFX] Texture asset path " ] = None ,
mesh_path : Annotated [ str | None , " [VFX] Mesh asset path " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
gradient : Annotated [ Any ,
" [VFX/Line/Trail] Gradient { colorKeys, alphaKeys} or { startColor, endColor} (dict or JSON string) " ] = None ,
curve : Annotated [ Any ,
" [VFX] Animation curve keys or { startValue, endValue} (array, dict, or JSON string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
event_name : Annotated [ str | None , " [VFX] Event name to send " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
velocity : Annotated [ Any ,
" [VFX] Event velocity [x,y,z] (array or JSON string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
size : Annotated [ Any , " [VFX] Event size (number or string) " ] = None ,
lifetime : Annotated [ Any , " [VFX] Event lifetime (number or string) " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
play_rate : Annotated [ Any ,
" [VFX] Playback speed multiplier (number or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
seed : Annotated [ Any , " [VFX] Random seed (integer or string) " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
reset_seed_on_play : Annotated [ Any ,
" [VFX] Reset seed on play (bool or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === LINE/TRAIL RENDERER PARAMETERS ===
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
positions : Annotated [ Any ,
" [Line] Positions [[x,y,z], ...] (array or JSON string) " ] = None ,
position : Annotated [ Any ,
" [Line/Trail] Single position [x,y,z] (array or JSON string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
index : Annotated [ Any , " [Line] Position index (integer or string) " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Width
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
width : Annotated [ Any ,
" [Line/Trail] Uniform width (number or string) " ] = None ,
start_width : Annotated [ Any ,
" [Line/Trail] Start width (number or string) " ] = None ,
end_width : Annotated [ Any ,
" [Line/Trail] End width (number or string) " ] = None ,
width_curve : Annotated [ Any ,
" [Line/Trail] Width curve (number or dict) " ] = None ,
width_multiplier : Annotated [ Any ,
" [Line/Trail] Width multiplier (number or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Color
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
color : Annotated [ Any ,
" [Line/Trail/VFX] Color [r,g,b,a] (array or JSON string) " ] = None ,
start_color_line : Annotated [ Any ,
" [Line/Trail] Start color (array or JSON string) " ] = None ,
end_color : Annotated [ Any ,
" [Line/Trail] End color (array or JSON string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Material & properties
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
material_path : Annotated [ str | None ,
" [Particle/Line/Trail] Material asset path " ] = None ,
trail_material_path : Annotated [ str | None ,
" [Particle] Trail material asset path " ] = None ,
loop : Annotated [ Any ,
" [Line] Connect end to start (bool or string) " ] = None ,
use_world_space : Annotated [ Any ,
" [Line] Use world space (bool or string) " ] = None ,
num_corner_vertices : Annotated [ Any ,
" [Line/Trail] Corner vertices (integer or string) " ] = None ,
num_cap_vertices : Annotated [ Any ,
" [Line/Trail] Cap vertices (integer or string) " ] = None ,
alignment : Annotated [ Literal [ " View " , " Local " , " TransformZ " ]
| None , " [Line/Trail] Alignment " ] = None ,
texture_mode : Annotated [ Literal [ " Stretch " , " Tile " , " DistributePerSegment " ,
" RepeatPerSegment " ] | None , " [Line/Trail] Texture mode " ] = None ,
generate_lighting_data : Annotated [ Any ,
" [Line/Trail] Generate lighting data for GI (bool or string) " ] = None ,
sorting_order : Annotated [ Any ,
" [Line/Trail/Particle] Sorting order (integer or string) " ] = None ,
sorting_layer_name : Annotated [ str | None ,
" [Renderer] Sorting layer name " ] = None ,
sorting_layer_id : Annotated [ Any ,
" [Renderer] Sorting layer ID (integer or string) " ] = None ,
render_mode : Annotated [ str | None ,
" [Particle] Render mode (Billboard, Stretch, HorizontalBillboard, VerticalBillboard, Mesh, None) " ] = None ,
sort_mode : Annotated [ str | None ,
" [Particle] Sort mode (None, Distance, OldestInFront, YoungestInFront, Depth) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === RENDERER COMMON PROPERTIES (Shadows, Lighting, Probes) ===
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
shadow_casting_mode : Annotated [ Literal [ " Off " , " On " , " TwoSided " ,
" ShadowsOnly " ] | None , " [Renderer] Shadow casting mode " ] = None ,
receive_shadows : Annotated [ Any ,
" [Renderer] Receive shadows (bool or string) " ] = None ,
shadow_bias : Annotated [ Any ,
" [Renderer] Shadow bias (number or string) " ] = None ,
light_probe_usage : Annotated [ Literal [ " Off " , " BlendProbes " , " UseProxyVolume " ,
" CustomProvided " ] | None , " [Renderer] Light probe usage mode " ] = None ,
reflection_probe_usage : Annotated [ Literal [ " Off " , " BlendProbes " , " BlendProbesAndSkybox " ,
" Simple " ] | None , " [Renderer] Reflection probe usage mode " ] = None ,
motion_vector_generation_mode : Annotated [ Literal [ " Camera " , " Object " ,
" ForceNoMotion " ] | None , " [Renderer] Motion vector generation mode " ] = None ,
rendering_layer_mask : Annotated [ Any ,
" [Renderer] Rendering layer mask for SRP (integer or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === PARTICLE RENDERER SPECIFIC ===
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
min_particle_size : Annotated [ Any ,
" [Particle] Min particle size relative to viewport (number or string) " ] = None ,
max_particle_size : Annotated [ Any ,
" [Particle] Max particle size relative to viewport (number or string) " ] = None ,
length_scale : Annotated [ Any ,
" [Particle] Length scale for stretched billboard (number or string) " ] = None ,
velocity_scale : Annotated [ Any ,
" [Particle] Velocity scale for stretched billboard (number or string) " ] = None ,
camera_velocity_scale : Annotated [ Any ,
" [Particle] Camera velocity scale for stretched billboard (number or string) " ] = None ,
normal_direction : Annotated [ Any ,
" [Particle] Normal direction 0-1 (number or string) " ] = None ,
pivot : Annotated [ Any ,
" [Particle] Pivot offset [x,y,z] (array or JSON string) " ] = None ,
flip : Annotated [ Any ,
" [Particle] Flip [x,y,z] (array or JSON string) " ] = None ,
allow_roll : Annotated [ Any ,
" [Particle] Allow roll for mesh particles (bool or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Shape creation (line_create_*)
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
start : Annotated [ Any ,
" [Line] Start point [x,y,z] (array or JSON string) " ] = None ,
end : Annotated [ Any ,
" [Line] End point [x,y,z] (array or JSON string) " ] = None ,
center : Annotated [ Any ,
" [Line] Circle/arc center [x,y,z] (array or JSON string) " ] = None ,
segments : Annotated [ Any ,
" [Line] Number of segments (integer or string) " ] = None ,
normal : Annotated [ Any ,
" [Line] Normal direction [x,y,z] (array or JSON string) " ] = None ,
start_angle : Annotated [ Any ,
" [Line] Arc start angle degrees (number or string) " ] = None ,
end_angle : Annotated [ Any ,
" [Line] Arc end angle degrees (number or string) " ] = None ,
control_point1 : Annotated [ Any ,
" [Line] Bezier control point 1 (array or JSON string) " ] = None ,
control_point2 : Annotated [ Any ,
" [Line] Bezier control point 2 (cubic) (array or JSON string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Trail specific
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
min_vertex_distance : Annotated [ Any ,
" [Trail] Min vertex distance (number or string) " ] = None ,
autodestruct : Annotated [ Any ,
" [Trail] Destroy when finished (bool or string) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
emitting : Annotated [ Any , " [Trail] Is emitting (bool or string) " ] = None ,
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Common vector params for shape/velocity
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
x : Annotated [ Any ,
" [Particle] Velocity X (number or MinMaxCurve dict) " ] = None ,
y : Annotated [ Any ,
" [Particle] Velocity Y (number or MinMaxCurve dict) " ] = None ,
z : Annotated [ Any ,
" [Particle] Velocity Z (number or MinMaxCurve dict) " ] = None ,
speed_modifier : Annotated [ Any ,
" [Particle] Speed modifier (number or MinMaxCurve dict) " ] = None ,
space : Annotated [ Literal [ " Local " , " World " ] |
None , " [Particle] Velocity space " ] = None ,
separate_axes : Annotated [ Any ,
" [Particle] Separate XYZ axes (bool or string) " ] = None ,
size_over_lifetime : Annotated [ Any ,
" [Particle] Size over lifetime (number or MinMaxCurve dict) " ] = None ,
size_x : Annotated [ Any ,
" [Particle] Size X (number or MinMaxCurve dict) " ] = None ,
size_y : Annotated [ Any ,
" [Particle] Size Y (number or MinMaxCurve dict) " ] = None ,
size_z : Annotated [ Any ,
" [Particle] Size Z (number or MinMaxCurve dict) " ] = None ,
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
) - > dict [ str , Any ] :
""" Unified VFX management tool. """
# Normalize action to lowercase to match Unity-side behavior
action_normalized = action . lower ( )
# Validate action against known actions using normalized value
if action_normalized not in ALL_ACTIONS :
# Provide helpful error with closest matches by prefix
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
prefix = action_normalized . split (
" _ " ) [ 0 ] + " _ " if " _ " in action_normalized else " "
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
available_by_prefix = {
" particle_ " : PARTICLE_ACTIONS ,
" vfx_ " : VFX_ACTIONS ,
" line_ " : LINE_ACTIONS ,
" trail_ " : TRAIL_ACTIONS ,
}
suggestions = available_by_prefix . get ( prefix , [ ] )
if suggestions :
return {
" success " : False ,
" message " : f " Unknown action ' { action } ' . Available { prefix } * actions: { ' , ' . join ( suggestions ) } " ,
}
else :
return {
" success " : False ,
" message " : (
f " Unknown action ' { action } ' . Use prefixes: "
" particle_*, vfx_*, line_*, trail_*. Run with action= ' ping ' to test connection. "
) ,
}
unity_instance = get_unity_instance_from_context ( ctx )
# Build parameters dict with normalized action to stay consistent with Unity
params_dict : dict [ str , Any ] = { " action " : action_normalized }
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Target
if target is not None :
params_dict [ " target " ] = target
if search_method is not None :
params_dict [ " searchMethod " ] = search_method
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === PARTICLE SYSTEM ===
# Pass through all values - C# side handles parsing (ParseColor, ParseVector3, ParseMinMaxCurve, ToObject<T>)
if duration is not None :
params_dict [ " duration " ] = duration
if looping is not None :
params_dict [ " looping " ] = looping
if prewarm is not None :
params_dict [ " prewarm " ] = prewarm
if start_delay is not None :
params_dict [ " startDelay " ] = start_delay
if start_lifetime is not None :
params_dict [ " startLifetime " ] = start_lifetime
if start_speed is not None :
params_dict [ " startSpeed " ] = start_speed
if start_size is not None :
params_dict [ " startSize " ] = start_size
if start_rotation is not None :
params_dict [ " startRotation " ] = start_rotation
if start_color is not None :
params_dict [ " startColor " ] = start_color
if gravity_modifier is not None :
params_dict [ " gravityModifier " ] = gravity_modifier
if simulation_space is not None :
params_dict [ " simulationSpace " ] = simulation_space
if scaling_mode is not None :
params_dict [ " scalingMode " ] = scaling_mode
if play_on_awake is not None :
params_dict [ " playOnAwake " ] = play_on_awake
if max_particles is not None :
params_dict [ " maxParticles " ] = max_particles
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Emission
if rate_over_time is not None :
params_dict [ " rateOverTime " ] = rate_over_time
if rate_over_distance is not None :
params_dict [ " rateOverDistance " ] = rate_over_distance
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Shape
if shape_type is not None :
params_dict [ " shapeType " ] = shape_type
if radius is not None :
params_dict [ " radius " ] = radius
if radius_thickness is not None :
params_dict [ " radiusThickness " ] = radius_thickness
if angle is not None :
params_dict [ " angle " ] = angle
if arc is not None :
params_dict [ " arc " ] = arc
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Noise
if strength is not None :
params_dict [ " strength " ] = strength
if frequency is not None :
params_dict [ " frequency " ] = frequency
if scroll_speed is not None :
params_dict [ " scrollSpeed " ] = scroll_speed
if damping is not None :
params_dict [ " damping " ] = damping
if octave_count is not None :
params_dict [ " octaveCount " ] = octave_count
if quality is not None :
params_dict [ " quality " ] = quality
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Module
if module is not None :
params_dict [ " module " ] = module
if enabled is not None :
params_dict [ " enabled " ] = enabled
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Burst
if time is not None :
params_dict [ " time " ] = time
if count is not None :
params_dict [ " count " ] = count
if min_count is not None :
params_dict [ " minCount " ] = min_count
if max_count is not None :
params_dict [ " maxCount " ] = max_count
if cycles is not None :
params_dict [ " cycles " ] = cycles
if interval is not None :
params_dict [ " interval " ] = interval
if probability is not None :
params_dict [ " probability " ] = probability
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Playback
if with_children is not None :
params_dict [ " withChildren " ] = with_children
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === VFX GRAPH ===
# Asset management parameters
if asset_name is not None :
params_dict [ " assetName " ] = asset_name
if folder_path is not None :
params_dict [ " folderPath " ] = folder_path
if template is not None :
params_dict [ " template " ] = template
if asset_path is not None :
params_dict [ " assetPath " ] = asset_path
if overwrite is not None :
params_dict [ " overwrite " ] = overwrite
if folder is not None :
params_dict [ " folder " ] = folder
if search is not None :
params_dict [ " search " ] = search
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Runtime parameters
if parameter is not None :
params_dict [ " parameter " ] = parameter
if value is not None :
params_dict [ " value " ] = value
if texture_path is not None :
params_dict [ " texturePath " ] = texture_path
if mesh_path is not None :
params_dict [ " meshPath " ] = mesh_path
if gradient is not None :
params_dict [ " gradient " ] = gradient
if curve is not None :
params_dict [ " curve " ] = curve
if event_name is not None :
params_dict [ " eventName " ] = event_name
if velocity is not None :
params_dict [ " velocity " ] = velocity
if size is not None :
params_dict [ " size " ] = size
if lifetime is not None :
params_dict [ " lifetime " ] = lifetime
if play_rate is not None :
params_dict [ " playRate " ] = play_rate
if seed is not None :
params_dict [ " seed " ] = seed
if reset_seed_on_play is not None :
params_dict [ " resetSeedOnPlay " ] = reset_seed_on_play
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# === LINE/TRAIL RENDERER ===
if positions is not None :
params_dict [ " positions " ] = positions
if position is not None :
params_dict [ " position " ] = position
if index is not None :
params_dict [ " index " ] = index
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Width
if width is not None :
params_dict [ " width " ] = width
if start_width is not None :
params_dict [ " startWidth " ] = start_width
if end_width is not None :
params_dict [ " endWidth " ] = end_width
if width_curve is not None :
params_dict [ " widthCurve " ] = width_curve
if width_multiplier is not None :
params_dict [ " widthMultiplier " ] = width_multiplier
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Color
if color is not None :
params_dict [ " color " ] = color
if start_color_line is not None :
params_dict [ " startColor " ] = start_color_line
if end_color is not None :
params_dict [ " endColor " ] = end_color
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Material & properties
if material_path is not None :
params_dict [ " materialPath " ] = material_path
if trail_material_path is not None :
params_dict [ " trailMaterialPath " ] = trail_material_path
if loop is not None :
params_dict [ " loop " ] = loop
if use_world_space is not None :
params_dict [ " useWorldSpace " ] = use_world_space
if num_corner_vertices is not None :
params_dict [ " numCornerVertices " ] = num_corner_vertices
if num_cap_vertices is not None :
params_dict [ " numCapVertices " ] = num_cap_vertices
if alignment is not None :
params_dict [ " alignment " ] = alignment
if texture_mode is not None :
params_dict [ " textureMode " ] = texture_mode
if generate_lighting_data is not None :
params_dict [ " generateLightingData " ] = generate_lighting_data
if sorting_order is not None :
params_dict [ " sortingOrder " ] = sorting_order
if sorting_layer_name is not None :
params_dict [ " sortingLayerName " ] = sorting_layer_name
if sorting_layer_id is not None :
params_dict [ " sortingLayerID " ] = sorting_layer_id
if render_mode is not None :
params_dict [ " renderMode " ] = render_mode
if sort_mode is not None :
params_dict [ " sortMode " ] = sort_mode
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Renderer common properties (shadows, lighting, probes)
if shadow_casting_mode is not None :
params_dict [ " shadowCastingMode " ] = shadow_casting_mode
if receive_shadows is not None :
params_dict [ " receiveShadows " ] = receive_shadows
if shadow_bias is not None :
params_dict [ " shadowBias " ] = shadow_bias
if light_probe_usage is not None :
params_dict [ " lightProbeUsage " ] = light_probe_usage
if reflection_probe_usage is not None :
params_dict [ " reflectionProbeUsage " ] = reflection_probe_usage
if motion_vector_generation_mode is not None :
params_dict [ " motionVectorGenerationMode " ] = motion_vector_generation_mode
if rendering_layer_mask is not None :
params_dict [ " renderingLayerMask " ] = rendering_layer_mask
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Particle renderer specific
if min_particle_size is not None :
params_dict [ " minParticleSize " ] = min_particle_size
if max_particle_size is not None :
params_dict [ " maxParticleSize " ] = max_particle_size
if length_scale is not None :
params_dict [ " lengthScale " ] = length_scale
if velocity_scale is not None :
params_dict [ " velocityScale " ] = velocity_scale
if camera_velocity_scale is not None :
params_dict [ " cameraVelocityScale " ] = camera_velocity_scale
if normal_direction is not None :
params_dict [ " normalDirection " ] = normal_direction
if pivot is not None :
params_dict [ " pivot " ] = pivot
if flip is not None :
params_dict [ " flip " ] = flip
if allow_roll is not None :
params_dict [ " allowRoll " ] = allow_roll
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Shape creation
if start is not None :
params_dict [ " start " ] = start
if end is not None :
params_dict [ " end " ] = end
if center is not None :
params_dict [ " center " ] = center
if segments is not None :
params_dict [ " segments " ] = segments
if normal is not None :
params_dict [ " normal " ] = normal
if start_angle is not None :
params_dict [ " startAngle " ] = start_angle
if end_angle is not None :
params_dict [ " endAngle " ] = end_angle
if control_point1 is not None :
params_dict [ " controlPoint1 " ] = control_point1
if control_point2 is not None :
params_dict [ " controlPoint2 " ] = control_point2
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Trail specific
if min_vertex_distance is not None :
params_dict [ " minVertexDistance " ] = min_vertex_distance
if autodestruct is not None :
params_dict [ " autodestruct " ] = autodestruct
if emitting is not None :
params_dict [ " emitting " ] = emitting
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Velocity/size axes
if x is not None :
params_dict [ " x " ] = x
if y is not None :
params_dict [ " y " ] = y
if z is not None :
params_dict [ " z " ] = z
if speed_modifier is not None :
params_dict [ " speedModifier " ] = speed_modifier
if space is not None :
params_dict [ " space " ] = space
if separate_axes is not None :
params_dict [ " separateAxes " ] = separate_axes
if size_over_lifetime is not None :
params_dict [ " size " ] = size_over_lifetime
if size_x is not None :
params_dict [ " sizeX " ] = size_x
if size_y is not None :
params_dict [ " sizeY " ] = size_y
if size_z is not None :
params_dict [ " sizeZ " ] = size_z
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Remove None values
params_dict = { k : v for k , v in params_dict . items ( ) if v is not None }
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
# Send to Unity
result = await send_with_unity_instance (
async_send_command_with_retry ,
unity_instance ,
" manage_vfx " ,
params_dict ,
)
v9 pre-release pruning (#528)
* refactor: Split ParseColorOrDefault into two overloads and change default to Color.white
* Auto-format Python code
* Remove unused Python module
* Refactored VFX functionality into multiple files
Tested everything, works like a charm
* Rename ManageVfx folder to just Vfx
We know what it's managing
* Clean up whitespace on plugin tools and resources
* Make ManageGameObject less of a monolith by splitting it out into different files
* Remove obsolete FindObjectByInstruction method
We also update the namespace for ManageVFX
* refactor: Consolidate editor state resources into single canonical implementation
Merged EditorStateV2 into EditorState, making get_editor_state the canonical resource. Updated Unity C# to use EditorStateCache directly. Enhanced Python implementation with advice/staleness enrichment, external changes detection, and instance ID inference. Removed duplicate EditorStateV2 resource and legacy fallback mapping.
* Validate editor state with Pydantic models in both C# and Python
Added strongly-typed Pydantic models for EditorStateV2 schema in Python and corresponding C# classes with JsonProperty attributes. Updated C# to serialize using typed classes instead of anonymous objects. Python now validates the editor state payload before returning it, catching schema mismatches early.
* Consolidate run_tests and run_tests_async into single async implementation
Merged run_tests_async into run_tests, making async job-based execution the default behavior. Removed synchronous blocking test execution. Updated RunTests.cs to start test jobs immediately and return job_id for polling. Changed TestJobManager methods to internal visibility. Updated README to reflect single run_tests_async tool. Python implementation now uses async job pattern exclusively.
* Validate test job responses with Pydantic models in Python
* Change resources URI from unity:// to mcpforunity://
It should reduce conflicts with other Unity MCPs that users try, and to comply with Unity's requests regarding use of their company and product name
* Update README with all tools + better listing for resources
* Update other references to resources
* Updated translated doc - unfortunately I cannot verify
* Update the Chinese translation of the dev docks
* Change menu item from Setup Window to Local Setup Window
We now differentiate whether it's HTTP local or remote
* Fix URIs for menu items and tests
* Shouldn't have removed it
* Minor edits from CodeRabbit feedback
* Don't use reflection which takes longer
* Fix failing python tests
* Add serialization helpers for ParticleSystem curves and MinMaxCurve types
Added SerializeAnimationCurve and SerializeMinMaxCurve helper methods to properly serialize Unity's curve types. Updated GetInfo to use these helpers for startLifetime, startSpeed, startSize, gravityModifier, and rateOverTime instead of only reading constant values.
* Use ctx param
* Update Server/src/services/tools/run_tests.py
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* Minor fixes
* Rename anything EditorStateV2 to just EditorState
It's the default, there's no old version
* Make infer_single_instance_id public by removing underscore prefix
* Fix Python tests, again
* Replace AI generated .meta files with actual Unity ones
* ## Pre-Launch Enhancements: Testing Infrastructure & Tool Improvements (#8)
* Add local test harness for fast developer iteration
Scripts for running the NL/T/GO test suites locally against a GUI Unity
Editor, complementing the CI workflows in .github/workflows/.
Benefits:
- 10-100x faster than CI (no Docker startup)
- Real-time Unity console debugging
- Single test execution for rapid iteration
- Auto-detects HTTP vs stdio transport
Usage:
./scripts/local-test/setup.sh # One-time setup
./scripts/local-test/quick-test.sh NL-0 # Run single test
./scripts/local-test/run-nl-suite-local.sh # Full suite
See scripts/local-test/README.md for details.
Also updated .gitignore to:
- Allow scripts/local-test/ to be tracked
- Ignore generated artifacts (reports/*.xml, .claude/local/, .unity-mcp/)
* Fix issue #525: Save dirty scenes for all test modes
Move SaveDirtyScenesIfNeeded() call outside the PlayMode conditional
so EditMode tests don't get blocked by Unity's "Save Scene" modal dialog.
This prevents MCP from timing out when running EditMode tests with unsaved
scene changes.
* fix: add missing FAST_FAIL_TIMEOUT constant in PluginHub
The FAST_FAIL_TIMEOUT class attribute was referenced on line 149 but never
defined, causing AttributeError on every ping attempt. This error was silently
caught by the broad 'except Exception' handler, causing all fast-fail commands
(read_console, get_editor_state, ping) to fail after 6 seconds of retries with
'ping not answered' error.
Added FAST_FAIL_TIMEOUT = 10 to define a 10-second timeout for fast-fail
commands, matching the intent of the existing fast-fail infrastructure.
* feat(ScriptableObject): enhance dry-run validation for AnimationCurve and Quaternion
Dry-run validation now validates value formats, not just property existence:
- AnimationCurve: Validates structure ({keys:[...]} or direct array), checks
each keyframe is an object, validates numeric fields (time, value, inSlope,
outSlope, inWeight, outWeight) and integer fields (weightedMode)
- Quaternion: Validates array length (3 for Euler, 4 for raw) or object
structure ({x,y,z,w} or {euler:[x,y,z]}), ensures all components are numeric
Refactored shared validation helpers into appropriate locations:
- ParamCoercion: IsNumericToken, ValidateNumericField, ValidateIntegerField
- VectorParsing: ValidateAnimationCurveFormat, ValidateQuaternionFormat
Added comprehensive XML documentation clarifying keyframe field defaults
(all default to 0 except as noted).
Added 5 new dry-run validation tests covering valid and invalid formats
for both AnimationCurve and Quaternion properties.
* test: fix integration tests after merge
- test_refresh_unity_retry_recovery: Mock now handles both refresh_unity and
get_editor_state commands (refresh_unity internally calls get_editor_state
when wait_for_ready=True)
- test_run_tests_async_forwards_params: Mock response now includes required
'mode' field for RunTestsStartResponse Pydantic validation
- test_get_test_job_forwards_job_id: Updated to handle GetTestJobResponse as
Pydantic model instead of dict (use model_dump() for assertions)
* Update warning message to apply to all test modes
Follow-up to PR #527: Since SaveDirtyScenesIfNeeded() now runs for all test modes, update the warning message to say 'tests' instead of 'PlayMode tests'.
* feat(run_tests): add wait_timeout to get_test_job to avoid client loop detection
When polling for test completion, MCP clients like Cursor can detect the
repeated get_test_job calls as 'looping' and terminate the agent.
Added wait_timeout parameter that makes the server wait internally for tests
to complete (polling Unity every 2s) before returning. This dramatically
reduces client-side tool calls from 10-20 down to 1-2, avoiding loop detection.
Usage: get_test_job(job_id='xxx', wait_timeout=30)
- Returns immediately if tests complete within timeout
- Returns current status if timeout expires (client can call again)
- Recommended: 30-60 seconds
* fix: use Pydantic attribute access in test_run_tests_async for merge compatibility
* revert: remove local test harness - will be submitted in separate PR
---------
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: dsarno <david@lighthaus.us>
Co-authored-by: Scott Jennings <scott.jennings+CIGINT@cloudimperiumgames.com>
2026-01-08 06:51:51 +08:00
[FEATURE]: Manage VFX function (#520)
Feature: Add ManageVFX
Add ManageVFX to the current function, support:
1. Modify LineRender, TrailRender properties
2. Modify particle system (legacy) properties
3. Modify VisualEffectGraph and use templates (does not support edit graph yet)
* Update Server/src/services/tools/utils.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update Server/src/services/tools/manage_vfx.py
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Update ManageVFX.cs
Added 6 helper methods :
ApplyCommonRendererProperties - Handles shadows, lighting, probes, sorting, rendering layer (used by ParticleSetRenderer, LineSetProperties, TrailSetProperties)
GetCommonRendererInfo - Returns common renderer info for GetInfo methods
ApplyWidthProperties - Generic width handling (used by LineSetWidth, TrailSetWidth)
ApplyColorProperties - Generic color handling with fadeEndAlpha option (used by LineSetColor, TrailSetColor)
SetRendererMaterial - Generic material assignment (used by LineSetMaterial, TrailSetMaterial)
ApplyLineTrailProperties - Shared Line/Trail properties like alignment, textureMode, numCornerVertices (used by LineSetProperties, TrailSetProperties)
* Optimize the code structure
Declutter the redundant files by adding RendererHelpers and VectorParsing.cs
* Minor Fixes
Minor Fixes based on AI-feedback
* Fixes
---------
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
2026-01-07 14:47:32 +08:00
return result if isinstance ( result , dict ) else { " success " : False , " message " : str ( result ) }