A Rust-based GraphQL and RPC backend framework with support for PostgreSQL, Deno, WebAssembly, and multiple deployment targets (Actix, AWS Lambda, Cloudflare Workers). Contains MCP (Model Context Protocol) router and bridge components.
Scoring was not performed
Dynamic schema embedding in tool description violates token efficiency and description length best practices. The execute_query description includes a full GraphQL introspection schema string, making the description size variable and potentially huge, wasting context window tokens.
No response schemas documented for either tool. LLMs cannot predict output structure, forcing them to guess what fields are returned and plan downstream tool calls blindly. This violates the pattern requirement to document return types.
execute_query 'variables' parameter is typed as 'object' with empty properties ({}). No description explains what variables are valid. LLMs cannot infer which variables a given query accepts without seeing the query text itself, forcing trial-and-error.
introspection tool description is minimal (80 chars) and lacks WHEN/WHY guidance. It does not explain the structure of the returned schema, making it hard for LLMs to understand what fields to extract or how to use the result.
execute_query 'query' parameter lacks a description. LLMs do not know if they should pass raw GraphQL strings, formatted SDL, or some other representation.
No guidance on error recovery. Error messages should tell LLMs what to do next (e.g., 'Invalid query syntax, check syntax in the query parameter' or 'Variable mismatch, call introspection first to see available queries'). Actionable errors are critical for agent self-correction.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 0 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 25 | - | v1 |