SNCF Open Data (ressources.data.sncf.com) — OpenDataSoft MCP. Open data for French national railways (SNCF) via the OpenDataSoft Explore v2.1 API: train schedules, stations, punctuality, ridership & the rail network. Keyless.
Three tools with complete schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. All tools have action verbs (search_, dataset_, query) and non-empty descriptions (avg 156 chars, within baseline 194). However, parameter descriptions are sparse or missing context. No output schemas documented. Error handling is minimal. The server follows basic tool definition structure but lacks LLM-optimization patterns for parameter constraints and recovery guidance.
Get metadata for one SNCF Open Data dataset (fields/schema, themes, record count) — call before query to learn the column names.
Query records from a SNCF Open Data dataset with ODSQL. Filter (where), aggregate (group_by/select), sort (order_by), paginate (limit/offset).
Search SNCF Open Data for datasets by keyword (train schedules, stations, punctuality, ridership & the rail network). Returns dataset_ids (pass to query/dataset_info), titles, themes and record counts.
Output schemas not documented. LLMs cannot predict response structure, forcing them to guess field names for downstream processing. search_datasets, dataset_info, and query all lack documented return types.
Parameter descriptions lack actionable constraints. 'limit' and 'offset' state ranges (1-100, default 20) but 'query', 'where', 'select', 'group_by', 'order_by' have no format guidance, examples, or ODSQL syntax hints. LLMs will hallucinate invalid ODSQL.
No error recovery guidance. callTool throws generic errors ('Unknown tool', 'Required argument missing', 'SNCF Open Data: 404') without suggesting next steps. LLMs cannot self-correct or retry intelligently.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2026-07-28+ | v2 |
Parameter 'query' in search_datasets and query tool lacks clarity on scope. Does 'query' search all fields or only titles? Does it support ODSQL syntax or free-text only? Ambiguity forces LLM guessing.
No pagination guidance in descriptions. While limit/offset are present, descriptions do not explain how to detect end-of-results or total record count. LLMs may not know when to stop paginating.