MCP Server to connect to Hugging Face spaces. Simple configuration, Claude Desktop friendly.
Critical failures across all dimensions. The server exposes a single tool 'dynamic_endpoint_tools' with a completely empty input schema (no properties defined). The description is vague and does not explain what the tool does, when to use it, or what it returns. Parameters are dynamically derived from Gradio API schemas at runtime, making static validation impossible. No parameter descriptions, no output schema documentation, and no error handling guidance. The tool name does not follow verb_noun convention and is plural/generic. This server cannot be safely used by LLMs without explicit per-endpoint schema documentation.
Dynamically generated tools from Hugging Face Space endpoints. Each endpoint from configured spaces becomes a tool with parameters derived from the Gradio API schema.
Input schema is completely empty: {"type":"object","properties":{},"description":"Schema varies per endpoint..."}. No parameters are defined, making it impossible for LLMs to understand what inputs the tool accepts or how to invoke it correctly.
Tool description is vague and incomplete: 'Dynamically generated tools from Hugging Face Space endpoints. Each endpoint from configured spaces becomes a tool with parameters derived from the Gradio API schema.' Does not explain WHAT the tool does, WHEN to use it, or WHAT it returns. Fails to meet 10-1024 character guideline with actionable content.
Tool name 'dynamic_endpoint_tools' does not follow verb_noun convention (e.g., 'call_endpoint', 'invoke_space_endpoint'). Plural form and generic 'tools' suffix obscure the actual action. LLMs cannot infer intent from the name alone.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 26 | 2026-07-28+ | v2 |
No output schema documented. The tool returns results from arbitrary Gradio endpoints, but there is no specification of what fields, types, or structure the LLM should expect. This forces LLMs to guess at response structure and breaks downstream tool chaining.
No error handling guidance. The endpoint_wrapper.ts catches errors and returns 'mcp-hfspace error: {message}' but provides no recovery hints, categorization (retryable vs fatal), or actionable next steps for the LLM.