MCP Language Server — Rust rewrite for million-line C/C++ codebase analysis. Exposes code context extraction capabilities to LLMs via the MCP protocol with support for multiple LSP backends (clangd, gopls, etc.)
Single tool 'echo' with minimal definition. Description is trivial (16 chars, below 20-char threshold). No output schema documented. Tool does not follow verb_noun naming (should be 'echo_message'). No error handling guidance. Fails multiple hard scoring rules: description under 20 chars, parameters lack descriptions, no output schema visible.
Echo back the input
Tool description 'Echo back the input' is 16 characters, below 20-char minimum. LLM cannot determine when/why to select this tool.
Input parameter 'message' has no description. LLM cannot infer whether this accepts plain text, JSON, code, or structured data.
No output schema documented. LLM cannot plan downstream tool calls or extract return data. Rubric requires 100% of A+ tools to have documented return types.
Tool name 'echo' does not follow verb_noun convention. Should be 'echo_message' to clarify action and object. Generic names like 'echo' or 'process' reduce LLM disambiguation.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 30 | <=2025-11-25 | v2 |
No error handling guidance. If echo fails (e.g., message too large, invalid encoding), LLM receives no recovery path.