Revit MCP has basic tool definitions with names and descriptions, but significant gaps in schema completeness, parameter documentation, and error handling. All 5 tools have descriptions (10-50 chars, below the 50-200 char optimum), but parameter descriptions are minimal or missing. Input schemas are present but lack detail (no enums, ranges, or format constraints). No output schemas documented. Error handling is absent, tools return raw strings with no recovery guidance. The architecture (STDIO pipe to Revit add-in) is sound, but the tool interface is underdeveloped for LLM reasoning.

• Document output schemas for all tools. E.g., get_wall_length should return {"wall_id": string, "length_feet": number, "length_meters": number} or {"error": string, "code": string}. This enables LLMs to chain tools and extract the right fields.
• Expand parameter descriptions to 50-100 chars with format and constraint details. E.g., 'wall_id: Revit element ID as a numeric string (e.g., "12345"). Call list_walls_by_level first if you only have a wall name.'
• Add error handling with recovery guidance. Return structured errors: {"success": false, "error": "File not found", "suggestion": "Check the file path or call list_available_documents() to see open files."}
• Implement input validation in the MCP server (RevitDocumentTools.cs). Validate file_path exists, wall_id is numeric, level_name matches known levels. Return actionable error messages, not stack traces.
• Add tool annotations (destructiveHint, readOnlyHint) to the schema. Mark open_revit_document and close_revit_document as destructive (they modify Revit state); mark get_* tools as readOnly. This helps LLMs reason about side effects.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry