Agent-Skills.md

Agent Skills: GitHub Actions Debugger

Debug and monitor GitHub Actions workflow runs. Check run status, view failed job logs, and troubleshoot CI failures. Use this when the user needs to investigate GitHub Actions failures, inspect job output, or identify the root cause of a broken workflow run.

UncategorizedID: opendatahub-io/ai-helpers/github-actions-debugger

Repository

opendatahub-io/ai-helpers
opendatahub-ioLicense: Apache-2.0
3674

Install this agent skill to your local

pnpm dlx add-skill https://github.com/opendatahub-io/ai-helpers/tree/HEAD/helpers/skills/github-actions-debugger

Skill Files

Browse the full folder contents for github-actions-debugger.

Download Skill

Loading file tree…

helpers/skills/github-actions-debugger/SKILL.md

Skill Metadata

Name
github-actions-debugger
Description
Debug and monitor GitHub Actions workflow runs. Check run status, view failed job logs, and troubleshoot CI failures. Use this when the user needs to investigate GitHub Actions failures, inspect job output, or identify the root cause of a broken workflow run.

GitHub Actions Debugger

This skill enables Claude to investigate GitHub Actions failures by:

  1. Listing recent workflow runs and their status
  2. Identifying failed jobs within a run
  3. Retrieving failed job logs
  4. Analyzing error messages and suggesting next debugging steps

Prerequisites

  • gh CLI installed and authenticated (gh auth status to verify)
  • Git repository with a GitHub remote configured (for automatic repo detection)

If gh is not installed, ask the user to follow installation instructions at https://cli.github.com/

If gh is not authenticated, ask the user to run gh auth login.

Instructions

When the user asks to check CI status, debug workflow failures, or view job logs, use the gh CLI commands below. All commands auto-detect the GitHub repository from the git remote in the current working directory. To target a different repository, add -R <owner>/<repo> to any command.

Handling GitHub Actions URLs

When the user provides a full GitHub Actions URL, parse it to extract the repository and run or job ID, then use -R to target that repository. Always use the repository from the URL, ignoring the local git remote.

Before constructing any command from a URL, validate the extracted values:

  • Owner names must only contain letters, digits, and hyphens.

  • Repository names must only contain letters, digits, hyphens, underscores, and dots.

  • Run IDs and job IDs must be purely numeric.

  • Workflow file names must only contain letters, digits, hyphens, underscores, and dots. If any extracted value fails its rule, reject it and ask the user to provide a corrected URL or ID.

  • Run URL format: https://github.com/<owner>/<repo>/actions/runs/<run-id>

    • Extract: -R <owner>/<repo> and <run-id>
    • Example: https://github.com/my-org/my-project/actions/runs/123456789 becomes gh run view 123456789 -R my-org/my-project

  • Workflow URL format: https://github.com/<owner>/<repo>/actions/workflows/<workflow-file>

    • Extract: -R <owner>/<repo> and --workflow <workflow-file>
    • Example: https://github.com/my-org/my-project/actions/workflows/ci.yaml becomes gh run list --workflow ci.yaml -R my-org/my-project

Commands

  1. List Recent Workflow Runs

    • Run gh run list --limit 10 to show the latest runs with their status, workflow name, and run ID
    • Add --status failure to filter for only failed runs: gh run list --status failure --limit 10

  2. Filter by Workflow File

    • Use --workflow to scope results to a specific workflow: gh run list --workflow ci.yaml --limit 10

  3. View a Specific Run

    • Use gh run view <run-id> to see the run summary including all jobs and their status: gh run view <run-id>
    • Add --log-failed to immediately stream the logs of all failed jobs: gh run view <run-id> --log-failed

  4. View Full Logs for a Run

    • To retrieve complete logs for every job in a run: gh run view <run-id> --log
    • Pipe through grep to isolate errors: gh run view <run-id> --log-failed 2>&1 | grep -i "error\|fail\|fatal" | head -50

  5. Watch a Running Workflow

    • To follow a currently active run in real time: gh run watch <run-id>

  6. Re-run Failed Jobs

    • If the user wants to retry only the failed jobs without re-running passing ones: gh run rerun <run-id> --failed

  7. Troubleshoot CI Failures

    • First, list recent failed runs to identify the run ID (step 1)
    • Then, view the run to see which jobs failed (step 3)
    • Then, retrieve the failed job logs (step 3 with --log-failed)
    • Analyze the log output to identify the root cause and suggest fixes

Examples

# List the 10 most recent runs across all workflows
gh run list --limit 10

# Show only failed runs for a specific workflow
gh run list --workflow ci.yaml --status failure --limit 10

# View run summary and failed job logs for run ID 123456789
gh run view 123456789 --log-failed

# Grep failed logs for actionable errors
gh run view 123456789 --log-failed 2>&1 | grep -i "error\|fail\|fatal" | head -50

# Re-run only the failed jobs
gh run rerun 123456789 --failed

Error Handling

  • If a run ID is not found, show the error and suggest running gh run list to find the correct ID
  • If the repository cannot be detected, suggest using -R <owner>/<repo> or running from within the correct git repository
  • If logs are unavailable (e.g., run too old or log retention expired), inform the user and suggest checking the GitHub Actions UI directly