github-workflow-automation

GitHub Workflow Automation

Patterns for automating GitHub workflows with AI assistance, inspired by Gemini CLI and modern DevOps practices.

When to Use This Skill

• Automating PR reviews with AI
• Setting up issue triage automation
• Creating GitHub Actions workflows
• Integrating AI into CI/CD pipelines
• Automating Git operations (rebases, cherry-picks)

Full YAML workflow files are in references/workflow_examples.md.

---

1. Automated PR Review

Create .github/workflows/ai-review.yml triggered on pull_request: [opened, synchronize].

Flow: checkout with fetch-depth: 0 → collect changed files and diff via git diff → send to an LLM (e.g., Claude) → post the response as a PR review comment.

Key elements:

• Permissions: pull-requests: write, contents: read
• Use actions/github-script@v7 to call the AI API and pulls.createReview
• Store the API key in secrets.ANTHROPIC_API_KEY

Review output should follow this structure: Summary → What looks good → Potential Issues → Suggestions → Security Notes.

For focused reviews, filter changed files with grep -E '\.(ts|tsx|js|jsx|py|go)$' before sending to the model.

---

2. Issue Triage Automation

Auto-label on Open

Create .github/workflows/issue-triage.yml triggered on issues: [opened].

Flow: read issue title & body → call AI with a classification prompt → apply labels (bug, enhancement, question, area labels) → if bug with no repro steps, post a comment asking for details.

Triage prompt returns JSON:

{
  "type": "bug | feature | question | docs | other",
  "severity": "low | medium | high | critical",
  "area": "frontend | backend | api | docs | ci | other",
  "summary": "one-line summary",
  "hasReproSteps": true,
  "suggestedLabels": [],
  "suggestedAssignees": []
}

Stale Issue Management

Use actions/stale@v9 on a daily cron. Mark issues stale after 60 days, close after 14 more. Exempt labels: pinned, security, in-progress.