Update to GTFS Validator

{
  "description": "Validate GTFS Schedule ZIP files or public GTFS feed URLs with GTFS Validator and explain prioritized findings.",
  "included_files": [],
  "name": "gtfs-validator",
  "skill_md_contents": "---\nname: gtfs-validator\ndescription: Validate GTFS Schedule ZIP files or public GTFS feed URLs with GTFS Validator and explain prioritized findings.\n---\n\n# GTFS Validator\n\nUse the connected GTFS Validator MCP server for GTFS Schedule validation.\n\n## Tool selection\n\n- Use `analyze_gtfs_file` for a user-uploaded GTFS Schedule ZIP.\n- Use `analyze_gtfs_url` for a public, directly downloadable GTFS ZIP URL.\n- Use `get_gtfs_rule` when the user asks what a rule ID means.\n\nPreserve the returned `status`, `analysis.input_mode`,\n`analysis.source_url_provided`, publishability, coverage, scores, severity\ncounts, feed metrics, and R9 ordering. Do not invent findings or claim that a\nfeed is publishable when the returned report says otherwise.\n\nTreat the Analyzer's scores, publishability verdict, severity counts, and R9\nordering as canonical. Do not recalculate or reinterpret them in the client.\nQuality- or Analytics-only findings do not make a feed invalid by themselves;\nfollow the returned publishability and blocker fields. Explain `partial`\nresults as incomplete coverage, and explain `fatal` results as an analysis\nfailure rather than a feed verdict.\n\nIf the server returns `FILE_TOO_LARGE`, explain that synchronous ChatGPT\nvalidation is bounded and provide the returned `analyzer_web_url` when present.\n"
}