Critical gaps across all three tools. Tool definitions are inferred from scattered Svelte/Rust files rather than centrally registered with explicit schemas. No input schemas visible in source code. Descriptions exist but are vague and lack LLM-optimization guidance. Parameters lack type definitions and constraints. No output schemas documented. Error handling absent. This server does not meet baseline production standards.

• Create a centralized MCP tool registry (e.g., in src-tauri/src/mcp.rs) with explicit tool definitions using schemars or similar. Each tool must have a name, description (50-200 chars), input schema with typed parameters, and output schema.
• Add JSON Schema definitions for all parameters. Example: load_toolpath.path should be {"type": "string", "description": "Absolute or relative file path to G-code file (e.g., /home/user/job.gcode). Must be readable by the server process."}
• Document output schemas. load_toolpath should return {"toolpath": Toolpath, "error": null} or {"toolpath": null, "error": string}. send_line should return {"success": bool, "ack": string, "error": string}.
• Expand descriptions to 50-200 chars with WHEN/WHY context. Example: 'send_line: Send a single G-code command to the CNC machine and wait for acknowledgment. Use for manual jogging or single-step execution. Returns ack string or error message.'
• Add error handling guidance. Document which errors are retryable (connection timeout) vs fatal (invalid G-code syntax). Provide recovery hints: 'If machine is offline, check USB connection and retry.'