Critical failure across all dimensions. No tool definitions are visible in the provided source code. The Python script documents tool name derivation but does not show actual MCP tool registrations with schemas, descriptions, or parameter definitions. The .mcp.json config only specifies the executable path and environment variables. Without access to the actual C# tool implementations in src/housecarl-mcp, no schema validation, description assessment, or parameter typing can be performed. This is a complete absence of verifiable definition quality.

• Add comprehensive tool descriptions (50-200 characters) to every tool. Each description must state WHAT the tool does, WHEN to use it, and any prerequisites. Example: 'housecarl_create' → 'Create a new plugin file with the specified name and initial configuration. Use this before adding records or scripts to a plugin.'
• Document input schemas with proper JSON Schema. Every parameter must have: name, type (string|number|boolean|array|object), description, and constraints (enum, pattern, min/max). Example: { 'plugin_name': { 'type': 'string', 'description': 'Name of the plugin (2-50 chars, alphanumeric + underscore)', 'pattern': '^[a-zA-Z0-9_]{2,50}$' } }
• Clarify vague tool names. Replace 'housecarl_create' with 'housecarl_create_plugin', 'housecarl_forward' with 'housecarl_forward_record', 'housecarl_place' with 'housecarl_place_object', 'housecarl_check' with 'housecarl_check_plugin_validity', and 'housecarl_skse' with 'housecarl_get_skse_info' or similar.
• Document output schemas for all tools. Specify the structure of returned data, including field names, types, and what each field represents. Include pagination info (limit, offset, total_count) for list tools.
• Add error handling guidance. For each tool, document what errors can occur (e.g., 'Plugin not found', 'Invalid file format'), what they mean, and what the LLM should do next (retry, ask user, try alternative tool).