MCP server for Grabblist wishlist and collection management service
CRITICAL: No tool definitions are visible in the source code. The validation script (scripts/validate_public_repo.py) only lists expected tool names in a set (EXPECTED_TOOLS) and validates that TOOLS.md documents them. The actual tool schemas, descriptions, input parameters, and output structures are NOT present in the provided source. The server is HTTP-based and points to https://mcp.grabbitapp.com/api/mcp, but the implementation details are hosted remotely and not inspectable. Per HARD SCORING RULES: 'If you cannot see the actual tool definition in the source (only inferred): cap that tool's overall at 50.' Since NO tool definitions are visible, only names are listed, all tools score 0 on schema and description. The validation script proves the server exists and exposes 23 tools, but provides zero evidence of definition quality.
No tool definitions visible in source code. Only tool names are listed in validation script; actual schemas, descriptions, parameters, and output structures are hosted remotely and not inspectable.
No input parameter schemas documented. Cannot verify parameter types, constraints, enums, or descriptions for any tool.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 21 | 2026-07-28+ | v2 |
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields without knowing response structure.
Destructive tools (delete_item, delete_collection) lack confirmation/dry-run patterns. No evidence of permission gates or audit logging.
No error handling guidance visible. Cannot verify recovery patterns, error classification, or actionable error messages.