MCP server for fetching and analyzing GitHub Actions workflow run logs and artifacts
Six tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. All tools are READ_ONLY except download_artifact (WRITE). Naming follows verb_noun convention (get_*, download_*). Descriptions exist but are minimal (10-200 chars) and lack context on when to use each tool or what they return. Parameter descriptions are present but sparse. No output schemas documented. Error handling guidance is absent. The server lacks idempotency hints, permission declarations, and recovery guidance, critical for agent reliability.
Download and extract an artifact
Get logs for a specific job. Note: This endpoint requires 'Actions: Read' permission on the repository. The logs endpoint may return a redirect to the actual log file location. For fine-grained PATs: Repository permission 'Actions' with at least 'Read' access. For classic PATs: 'repo' scope or 'actions:read' scope
Get artifacts for a workflow run
Get jobs for a workflow run
Get details of a specific workflow run
Get workflow runs for the repository
No output schemas documented. LLMs cannot predict what fields each tool returns, forcing them to guess at downstream field names and risk broken tool chains.
Descriptions are minimal and lack context. 'Get workflow runs for the repository' does not explain when to call this vs get_workflow_run, what fields are returned, or how pagination works.
No error handling guidance. If a job_id is invalid or logs are unavailable, the tool provides no recovery hint (e.g., 'Try get_workflow_jobs() to find valid job IDs').
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2026-07-28+ | v2 |
download_artifact is a WRITE operation but lacks idempotency hint or confirmation step. Agents may retry and create duplicate artifact extractions.
No permission declarations. Tools require 'Actions: Read' permission but this is not stated in tool metadata, preventing least-privilege agent configuration.