A backend Python application for managing MCP servers with Claude Desktop
Server has 6 tools with significant definition gaps. Tool names lack consistent verb-noun structure (e.g., 'restart_claude_desktop' vs 'restart_claude'). Descriptions are present but generic (avg ~80 chars, below 194-char baseline). Critical issue: set_server_enabled_status has required=[] but documents 'enabled' as required, schema contradicts description. Parameters lack type constraints (enums, ranges). No output schemas documented. Error handling guidance absent. Tools 5-6 (install_repo_mcp_server, install_local_mcp_server) are inferred from index.js reference only, cannot verify actual registration, capping those at 50.
Install an MCP server whose code is cloned locally on your computer
Register a new MCP server configuration. Tests the server command before registration.
Install an MCP server via npx or uvx
Restart the Claude application
Finds, terminates, and restarts the Claude Desktop application
Enable or disable an MCP server in Claude Desktop. Requires either server_id OR server_name (not both), and an enabled flag.
set_server_enabled_status has schema required=[] but description states 'enabled' is required. Schema contradicts documentation, causing LLM confusion about mandatory parameters.
No output schemas documented for any tool. LLMs cannot infer return structure, forcing them to guess what fields are available for downstream tool chaining.
Parameter descriptions lack format/constraint details. 'command' in install_mcp_server is array of strings but no guidance on expected format (e.g., 'absolute path to executable'). 'environment' object has no schema for key-value pairs.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 42 | <=2025-11-25 | v2 |
Duplicate/overlapping tool names: 'restart_claude_desktop' and 'restart_claude' both restart Claude but with unclear distinction. LLM cannot determine which to call.
No error handling guidance. Tools lack descriptions of failure modes, recovery steps, or actionable error messages. E.g., install_mcp_server says 'Tests the server command before registration' but does not document what happens if test fails.