MCP server for evaluating machine learning datasets using analysis tools and providing project recommendations
All 4 tools have minimal descriptions (10-20 chars) and lack parameter descriptions entirely. Input schemas are present but trivial (single string param). No output schemas documented. Error handling returns generic {'error': str(e)} with no recovery guidance. Tool names lack clear action verbs (analyze_dataset_tool, dataset_score_tool are verbose; score_dataset would be clearer). No tool annotations, no idempotency hints, no permission gates. Security validation exists (validate_dataset_path) but is not documented in tool descriptions. Composition is reasonable (4 focused tools) but descriptions do not explain when to use each vs. the others.
Return complete dataset analysis.
Return dataset score.
Generate report.
Return project ideas.
All tool descriptions are under 20 characters and provide no context for LLM selection. 'Return complete dataset analysis' and 'Return dataset score' do not explain when to call each tool, what they return, or dependencies.
No parameter descriptions. The 'dataset_path' parameter appears in all 4 tools but has no description explaining format, validation rules, or whether relative/absolute paths are accepted. LLMs cannot infer this from the name alone.
Output schemas are not documented. Tools return dicts with keys like 'dataset_score', 'project_recommendations', 'report', but LLMs have no schema to understand structure, field types, or what to extract for downstream calls.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 31 | 2026-07-28+ | v2 |
Error handling returns bare {'error': str(e)} with no recovery guidance. LLMs cannot determine if the error is retryable, user-fixable, or fatal. No actionable next steps provided.
Tool names are verbose and lack clear action verbs. 'analyze_dataset_tool' should be 'analyze_dataset'; 'dataset_score_tool' should be 'score_dataset'. Current names are 18-27 chars; production baseline is 18 chars average. Verbose names waste tokens and reduce clarity.