Lurkr is a static analysis scanner for MCP server configurations, not an MCP server itself. It analyzes fixture files containing tool declarations without schemas, descriptions, or parameter definitions. The 8 tools listed are test fixtures extracted from .mcp.json and .ts files with minimal metadata. Most tools lack descriptions (6/8), none have visible input schemas, and parameter definitions are absent. Tools like 'shell' and 'run' are marked DESTRUCTIVE but have no error handling, confirmation patterns, or permission gates. This is a security scanning tool, not a production MCP server, so evaluation against the tool quality rubric is not applicable in the traditional sense.

• Add comprehensive JSON Schema input definitions for all tools, including parameter types (string, number, boolean, array, object), required fields, and constraints (enum, minLength, pattern, minimum, maximum).
• Write full descriptions (50-200 chars) for every tool explaining WHAT it does, WHEN to use it, and WHAT it returns. Current descriptions are missing or trivial (e.g., 'd').
• For destructive tools (shell, run), implement confirmation patterns: add a 'dry_run' parameter and require explicit approval before execution. Document error recovery paths.
• Add parameter descriptions for all inputs. E.g., 'shell' tool should document: command (string, required), timeout (number, optional, default 30s), working_directory (string, optional).
• Document output schemas for all tools. Specify what fields are returned, their types, and what IDs/references are included for tool chaining.