A typed, Rust-first Model Context Protocol server for Blender. Provides typed Blender automation: modelling, materials, node graphs, animation, rigging, rendering and import/export.
Critical gaps across all dimensions. Only 5 tools total with minimal schema visibility. Tool names lack clear action verbs (blender.status, scene.summary, tools.categories.list are acceptable but generic). Descriptions are present but extremely brief (10-30 chars), well below the 50-200 char LLM-optimized baseline. Input schemas visible for only 2 of 5 tools; 3 tools (blender.status, scene.summary, tools.categories.list, object.create) have no visible parameter definitions in source. No output schemas documented. No parameter descriptions. No error handling guidance. No composition hints for multi-step workflows. The codebase shows sophisticated Rust architecture (workspace, schemars integration, rmcp framework) but tool definitions are not exposed in the provided source excerpts.
Check Blender connection status and capabilities
Create a new object in the Blender scene
Get a summary of what is in the current Blender scene
Enable a tool category to make its tools available
List all available tool categories and their enabled status
No input schemas visible for 3 of 5 tools (blender.status, scene.summary, object.create).
All tool descriptions are 10-30 characters, far below the 50-200 char LLM-optimized baseline. Descriptions lack WHAT/WHEN/WHY context needed for tool selection.
No parameter descriptions visible for any tool. tools.categories.enable has a 'category' param with type string but no description explaining what category identifiers are valid or how to discover them.
No output schemas documented. LLMs cannot plan downstream calls or extract required fields. No indication of what blender.status, scene.summary, or object.create return.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 36 | <=2025-11-25 | v2 |
No error handling guidance. No indication of retryability, user-fixable vs fatal errors, or recovery steps. Agents cannot self-correct on failures.