Loading ToolBench
MCP server that fetches and converts Wikipedia articles to Markdown using SSE transport
Single tool with clear naming (read_wikipedia_article) and functional description. Input schema present with type and description. However, output schema is undocumented (returns string with no structure guidance), and parameter validation is minimal. Error handling exists but lacks recovery guidance for LLMs. Tool follows basic patterns but lacks production-grade polish expected for A/B tier.
Fetch a Wikipedia article at the provided URL, parse its main content, convert it to Markdown, and return the resulting text.
Output schema undocumented. Tool returns a string (markdown text) with no structure guidance. LLMs cannot plan downstream operations or validate response format.
Parameter 'url' lacks format constraints. Description says 'must start with http or https' but no regex pattern, length limits, or examples. LLMs may pass malformed URLs.
Error messages lack recovery guidance. When content div not found or HTTP fails, errors state the problem but do not suggest next steps (e.g., 'Try a different Wikipedia URL' or 'Verify the article exists').
No timeout or rate-limit documentation. Tool calls requests.get() with timeout=10 but does not document this constraint or what happens if Wikipedia is slow/unavailable.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2026-07-28+ | v2 |
Tool description lacks 'when to use' context. Does not explain when to call this vs. other information retrieval methods, or what types of Wikipedia articles it handles best.