MCP server that feeds AI tools the currently selected Unity GameObjects and assets of interest. Requires the "Unity AI Context Bridge" Unity package to be installed in the Unity project.
Single tool with clear naming (get_unity_context_items) and adequate description (194 chars). Input schema present with type and description. Tool annotations (readOnlyHint) correctly applied. However, output schema is not formally documented, only inferred from code. Parameter description is minimal (13 chars: 'URL-encoded Unity project full path'). Error handling exists but lacks recovery guidance for LLMs. No pagination, batch operations, or composition patterns evident.
Returns information about the Unity GameObjects, assets, and other editor items user selected in the Unity editor that are relevant to the active AI task.
Output schema not formally documented. Code returns {content: [{type, text}], structuredContent: unityContext} but LLMs cannot see this structure in tool definition. Agents cannot plan downstream operations or extract specific fields.
Parameter description too brief (13 chars). 'URL-encoded Unity project full path' lacks context on format, constraints, or examples. Should explain: expected format (absolute path), encoding requirement, and what happens if invalid.
Error messages in code are informative (e.g., 'projectPath must be URL encoded') but tool description does not document failure modes or recovery steps. LLMs cannot infer what to do if the call fails.
Tool depends on external Unity Editor Extension (bridge.json at Temp/unity-ai-context-bridge/bridge.json). This dependency is not documented in tool description. LLMs cannot know the prerequisite or diagnose missing setup.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | 2025-06-18+ | v2 |
No pagination or result limiting documented. If Unity context returns hundreds of GameObjects/assets, response could exhaust context window. Tool description should state expected result size and any limits.