Loading ToolBench
Generates API reference documentation from codebases
This MCP server has a single tool with severely incomplete specification. The tool definition is visible in client.py but lacks essential quality attributes required for production use. The description is minimal (51 chars), parameter descriptions are absent, there is no documented output schema, no error handling guidance, and the implementation appears to be a mock rather than functional. The server would not be usable in a production agent context due to missing schema details and lack of error recovery patterns.
Generate API reference documentation from a codebase
Parameter descriptions completely missing. 'codebase_path' and 'doc_type' have no descriptions explaining their expected format, valid values, constraints, or examples.
No output schema documented. The tool returns a dict with 'markdown_content' and 'extracted_endpoints' but these fields are undocumented. LLMs cannot plan downstream usage or extract the right data without a formal return schema.
Tool description is too short (51 chars) and lacks context for selection. Does not explain WHEN to use this tool, any prerequisites, or how it differs from other documentation tools.
No error handling or recovery guidance. The implementation has no validation, no error classification, and no actionable error messages. If the codebase_path is invalid or doc_type is unrecognized, there is no guidance for the LLM to self-correct.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 32 | 2026-07-28+ | v2 |
Mock implementation, not functional. The generate_docs method returns hardcoded endpoints instead of actually parsing the provided codebase. This is a placeholder, not a working tool.
Parameter 'doc_type' has no enum constraint. Accepts free-form strings with no validation. Should define valid documentation types (api_reference, architecture, changelog, etc.) as an enum to prevent hallucinated values.
Parameter 'codebase_path' lacks format constraints. No guidance on whether it should be absolute or relative, whether it supports glob patterns, or how the tool handles missing paths. Invites invalid input from LLMs.