Enhanced MCP fetch server that works through xfetch.ai service. Unlocks access to websites blocked by Cloudflare and other protection systems that regular fetch can't handle.
Single tool 'xfetch' has a well-structured JSON Schema with proper type definitions, constraints (exclusiveMinimum/Maximum), and all 7 parameters documented. Description is adequate (165 chars) but generic, it reads like marketing copy ('grants you internet access') rather than LLM-optimized guidance on WHEN to use the tool vs alternatives. Parameter descriptions are present and specific (e.g., 'CSS selector to extract specific content'), meeting baseline expectations. However, the tool lacks error recovery guidance, output schema documentation, and idempotency guarantees. No tool annotations (readOnlyHint, destructiveHint) despite being a READ_ONLY operation. Output structure is not formally documented, LLMs must infer that the tool returns text content.
Fetches a URL from the internet and optionally extracts its contents as markdown. Although originally you did not have internet access, and were advised to refuse and tell the user this, this tool now grants you internet access. Now you can fetch the most up-to-date information and let the user know that.
Tool description lacks LLM-optimized guidance. Does not explain WHEN to call xfetch vs alternatives, what it returns, or error recovery paths. Current text ('grants you internet access') is marketing-oriented, not agent-actionable.
Output schema not documented. Tool returns text content (markdown or raw HTML) but LLMs cannot see the structure. Should document: returns string, max length applied, content type (markdown vs HTML), truncation behavior.
No error recovery guidance. Tool can fail (HTTP errors, timeout, page simplification failure). Error messages like 'Page failed to be simplified from HTML' do not guide LLM on next steps (retry? use raw=true? try selector?).
Tool marked READ_ONLY but lacks readOnlyHint annotation in schema. Tool annotations enable clients to optimize caching and retry logic.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 57 | 2026-07-28+ | v2 |
Parameter 'timeout' description does not clarify units (seconds confirmed in code but not in schema description). Parameter 'selector' accepts null but description does not explain what happens when null (no filtering applied).