This is an educational repository, not a production MCP server. Tool definitions are minimal and scattered across multiple capstone projects. Only 5 tools are identifiable; 3 have basic descriptions but lack proper input schemas with type definitions. The read_file and run_shell tools have minimal parameter documentation. incidents_* tools have better descriptions but no visible input schema validation. No output schemas documented. Error handling is absent. This is educational skeleton code, not a polished MCP implementation.

• Add explicit JSON Schema to all tool inputs: define types (string, number, enum), constraints (minLength, maxLength, pattern, enum values), and required fields. Example: incidents_list.severity should be {type: 'string', enum: ['p0', 'p1', 'p2']}.
• Document output schemas for each tool. Specify what fields are returned, their types, and which fields are safe to pass to downstream tools. Example: incidents_list should return {incidents: [{id, severity, title, created_at}], total_count}.
• Expand tool descriptions to 50 - 200 characters with WHEN/WHY context. Example: 'Read a file from the sandbox directory. Use this to inspect code, logs, or config files. Returns up to 10KB; larger files are truncated.'
• Add parameter descriptions explaining format and constraints. Example: read_file.path should say 'Relative path from sandbox root (e.g., src/main.ts). Must not contain .. or absolute paths.'
• Implement error handling with recovery guidance. Example: 'File not found. Check the path and try list_files() to see available files.' or 'Command timed out after 30s. Try a simpler command or break the task into steps.'