MCP server for deterministic OpenDRIVE tooling with pageldev/libOpenDRIVE as the first-stage backend.
Server has 6 tools with explicit schemas and basic descriptions. All tools are READ_ONLY and follow verb_noun naming (summarize_, list_, get_). However, parameter descriptions are minimal or absent (e.g., 'Path to the OpenDRIVE file' repeated 5 times with no context on format/constraints). No output schemas documented. Error handling is minimal, tools return generic JSON responses without recovery guidance. Schemas are present but lack parameter descriptions beyond type declarations. No tool annotations (readOnlyHint, etc.). This is a functional but underdeveloped server typical of early-stage domain tools.
Summarize reference-line geometry segments for one road or all roads in an OpenDRIVE file.
Convert OpenDRIVE road coordinates to world coordinates.
List lanes for a selected road in an OpenDRIVE file.
List roads in an OpenDRIVE file through the selected backend.
Summarize an OpenDRIVE file through the selected backend.
Run basic libOpenDRIVE sanity checks on an OpenDRIVE file.
Parameter descriptions are missing or trivial. 'Path to the OpenDRIVE file' appears 5 times with no guidance on format (absolute vs relative), file size limits, or expected structure. LLMs cannot infer whether to pass '/path/to/file.xodr' or 'file.xodr'.
No output schemas documented. Tools return JSON responses (via model_dump) but LLMs have no visibility into the structure of 'ok', 'data', or error fields. This prevents downstream tool chaining and forces LLMs to guess field names.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). All tools are READ_ONLY but this is not declared in the schema. Clients cannot optimize caching or retry logic without explicit hints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | 2026-07-28+ | v2 |
Error handling returns generic JSON with 'ok' flag but no recovery guidance. When a file is not found or invalid, the response does not suggest next steps (e.g., 'Try list_roads() first to verify the file exists'). Errors are not categorized as retryable vs fatal.
Parameter 'road_id' in get_road_geometry_summary is optional but not marked as such in the schema (required array is ['file_path'] only). However, the description does not explain what happens when road_id is omitted (all roads? error?). Ambiguous behavior.