claude-code-action/docs/solutions.md

Solutions & Use Cases

This guide provides complete, ready-to-use solutions for common automation scenarios with Claude Code Action. Each solution includes working examples, configuration details, and expected outcomes.

📋 Table of Contents

• Automatic PR Code Review
• Review Only Specific File Paths
• Review PRs from External Contributors
• Custom PR Review Checklist
• Scheduled Repository Maintenance
• Issue Auto-Triage and Labeling
• Documentation Sync on API Changes
• Security-Focused PR Reviews

---

Automatic PR Code Review

When to use: Automatically review every PR opened or updated in your repository.

Basic Example (No Tracking)

name: Claude Auto Review
on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          fetch-depth: 1

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}

            Please review this pull request with a focus on:
            - Code quality and best practices
            - Potential bugs or issues
            - Security implications
            - Performance considerations

            Note: The PR branch is already checked out in the current working directory.

            Use `gh pr comment` for top-level feedback.
            Use `mcp__github_inline_comment__create_inline_comment` to highlight specific code issues.
            Only post GitHub comments - don't submit review text as messages.

          claude_args: |
            --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*)"

Key Configuration:

• Triggers on opened and synchronize (new commits)
• Always include REPO and PR NUMBER for context
• Specify tools for commenting and reviewing
• PR branch is pre-checked out

Expected Output: Claude posts review comments directly to the PR with inline annotations where appropriate.

Enhanced Example (With Progress Tracking)

Want visual progress tracking for PR reviews? Use track_progress: true to get tracking comments like in v0.x:

name: Claude Auto Review with Tracking
on:
  pull_request:
    types: [opened, synchronize, ready_for_review, reopened]

jobs:
  review:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          fetch-depth: 1

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          track_progress: true # ✨ Enables tracking comments
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}

            Please review this pull request with a focus on:
            - Code quality and best practices
            - Potential bugs or issues
            - Security implications
            - Performance considerations

            Provide detailed feedback using inline comments for specific issues.

          claude_args: |
            --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*)"

Benefits of Progress Tracking:

• Visual Progress Indicators: Shows "In progress" status with checkboxes
• Preserves Full Context: Automatically includes all PR details, comments, and attachments
• Migration-Friendly: Perfect for teams moving from v0.x who miss tracking comments
• Works with Custom Prompts: Your prompt becomes custom instructions while maintaining GitHub context

Expected Output:

1. Claude creates a tracking comment: "Claude Code is reviewing this pull request..."
2. Updates the comment with progress checkboxes as it works
3. Posts detailed review feedback with inline annotations
4. Updates tracking comment to "Completed" when done

---

Review Only Specific File Paths

When to use: Review PRs only when specific critical files change.

Complete Example:

name: Review Critical Files
on:
  pull_request:
    types: [opened, synchronize]
    paths:
      - "src/auth/**"
      - "src/api/**"
      - "config/security.yml"

jobs:
  security-review:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          fetch-depth: 1

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}

            This PR modifies critical authentication or API files.

            Please provide a security-focused review with emphasis on:
            - Authentication and authorization flows
            - Input validation and sanitization
            - SQL injection or XSS vulnerabilities
            - API security best practices

            Note: The PR branch is already checked out.

            Post detailed security findings as PR comments.

          claude_args: |
            --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*)"

Key Configuration:

• paths: filter triggers only for specific file changes
• Custom prompt emphasizes security for sensitive areas
• Useful for compliance or security reviews

Expected Output: Security-focused review when critical files are modified.

---

Review PRs from External Contributors

When to use: Apply stricter review criteria for external or new contributors.

Complete Example:

name: External Contributor Review
on:
  pull_request:
    types: [opened, synchronize]

jobs:
  external-review:
    if: github.event.pull_request.author_association == 'FIRST_TIME_CONTRIBUTOR'
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          fetch-depth: 1

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}
            CONTRIBUTOR: ${{ github.event.pull_request.user.login }}

            This is a first-time contribution from @${{ github.event.pull_request.user.login }}.

            Please provide a comprehensive review focusing on:
            - Compliance with project coding standards
            - Proper test coverage (unit and integration)
            - Documentation for new features
            - Potential breaking changes
            - License header requirements

            Be welcoming but thorough in your review. Use inline comments for code-specific feedback.

          claude_args: |
            --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr view:*)"

Key Configuration:

• if: condition targets specific contributor types
• Includes contributor username in context
• Emphasis on onboarding and standards

Expected Output: Detailed review helping new contributors understand project standards.

---

Custom PR Review Checklist

When to use: Enforce specific review criteria for your team's workflow.

Complete Example:

name: PR Review Checklist
on:
  pull_request:
    types: [opened, synchronize]

jobs:
  checklist-review:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          fetch-depth: 1

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}

            Review this PR against our team checklist:

            ## Code Quality
            - [ ] Code follows our style guide
            - [ ] No commented-out code
            - [ ] Meaningful variable names
            - [ ] DRY principle followed

            ## Testing
            - [ ] Unit tests for new functions
            - [ ] Integration tests for new endpoints
            - [ ] Edge cases covered
            - [ ] Test coverage > 80%

            ## Documentation
            - [ ] README updated if needed
            - [ ] API docs updated
            - [ ] Inline comments for complex logic
            - [ ] CHANGELOG.md updated

            ## Security
            - [ ] No hardcoded credentials
            - [ ] Input validation implemented
            - [ ] Proper error handling
            - [ ] No sensitive data in logs

            For each item, check if it's satisfied and comment on any that need attention.
            Post a summary comment with checklist results.

          claude_args: |
            --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*)"

Key Configuration:

• Structured checklist in prompt
• Systematic review approach
• Team-specific criteria

Expected Output: Systematic review with checklist results and specific feedback.

---

Scheduled Repository Maintenance

When to use: Regular automated maintenance tasks.

Complete Example:

name: Weekly Maintenance
on:
  schedule:
    - cron: "0 0 * * 0" # Every Sunday at midnight
  workflow_dispatch: # Manual trigger option

jobs:
  maintenance:
    runs-on: ubuntu-latest
    permissions:
      contents: write
      issues: write
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          fetch-depth: 0

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: |
            REPO: ${{ github.repository }}

            Perform weekly repository maintenance:

            1. Check for outdated dependencies in package.json
            2. Scan for security vulnerabilities using `npm audit`
            3. Review open issues older than 90 days
            4. Check for TODO comments in recent commits
            5. Verify README.md examples still work

            Create a single issue summarizing any findings.
            If critical security issues are found, also comment on open PRs.

          claude_args: |
            --allowedTools "Read,Bash(npm:*),Bash(gh issue:*),Bash(git:*)"

Key Configuration:

• schedule: for automated runs
• workflow_dispatch: for manual triggering
• Comprehensive tool permissions for analysis

Expected Output: Weekly maintenance report as GitHub issue.

---

Issue Auto-Triage and Labeling

When to use: Automatically categorize and prioritize new issues.

Complete Example:

name: Issue Triage
on:
  issues:
    types: [opened]

jobs:
  triage:
    runs-on: ubuntu-latest
    permissions:
      issues: write
      id-token: write
    steps:
      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: |
            REPO: ${{ github.repository }}
            ISSUE NUMBER: ${{ github.event.issue.number }}
            TITLE: ${{ github.event.issue.title }}
            BODY: ${{ github.event.issue.body }}
            AUTHOR: ${{ github.event.issue.user.login }}

            Analyze this new issue and:
            1. Determine if it's a bug report, feature request, or question
            2. Assess priority (critical, high, medium, low)
            3. Suggest appropriate labels
            4. Check if it duplicates existing issues

            Based on your analysis, add the appropriate labels using:
            `gh issue edit [number] --add-label "label1,label2"`

            If it appears to be a duplicate, post a comment mentioning the original issue.

          claude_args: |
            --allowedTools "Bash(gh issue:*),Bash(gh search:*)"

Key Configuration:

• Triggered on new issues
• Issue context in prompt
• Label management capabilities

Expected Output: Automatically labeled and categorized issues.

---

Documentation Sync on API Changes

When to use: Keep docs up-to-date when API code changes.

Complete Example:

name: Sync API Documentation
on:
  pull_request:
    types: [opened, synchronize]
    paths:
      - "src/api/**/*.ts"
      - "src/routes/**/*.ts"

jobs:
  doc-sync:
    runs-on: ubuntu-latest
    permissions:
      contents: write
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          ref: ${{ github.event.pull_request.head.ref }}
          fetch-depth: 0

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}

            This PR modifies API endpoints. Please:

            1. Review the API changes in src/api and src/routes
            2. Update API.md to document any new or changed endpoints
            3. Ensure OpenAPI spec is updated if needed
            4. Update example requests/responses

            Use standard REST API documentation format.
            Commit any documentation updates to this PR branch.

          claude_args: |
            --allowedTools "Read,Write,Edit,Bash(git:*)"

Key Configuration:

• Path-specific trigger
• Write permissions for doc updates
• Git tools for committing

Expected Output: API documentation automatically updated with code changes.

---

Security-Focused PR Reviews

When to use: Deep security analysis for sensitive repositories.

Complete Example:

name: Security Review
on:
  pull_request:
    types: [opened, synchronize]

jobs:
  security:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      security-events: write
      id-token: write
    steps:
      - uses: actions/checkout@v5
        with:
          fetch-depth: 1

      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          # Optional: Add track_progress: true for visual progress tracking during security reviews
          # track_progress: true
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}

            Perform a comprehensive security review:

            ## OWASP Top 10 Analysis
            - SQL Injection vulnerabilities
            - Cross-Site Scripting (XSS)
            - Broken Authentication
            - Sensitive Data Exposure
            - XML External Entities (XXE)
            - Broken Access Control
            - Security Misconfiguration
            - Cross-Site Request Forgery (CSRF)
            - Using Components with Known Vulnerabilities
            - Insufficient Logging & Monitoring

            ## Additional Security Checks
            - Hardcoded secrets or credentials
            - Insecure cryptographic practices
            - Unsafe deserialization
            - Server-Side Request Forgery (SSRF)
            - Race conditions or TOCTOU issues

            Rate severity as: CRITICAL, HIGH, MEDIUM, LOW, or NONE.
            Post detailed findings with recommendations.

          claude_args: |
            --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr diff:*)"

Key Configuration:

• Security-focused prompt structure
• OWASP alignment
• Severity rating system

Expected Output: Detailed security analysis with prioritized findings.

---

Tips for All Solutions

Always Include GitHub Context

prompt: |
  REPO: ${{ github.repository }}
  PR NUMBER: ${{ github.event.pull_request.number }}
  [Your specific instructions]

Common Tool Permissions

• PR Comments: Bash(gh pr comment:*)
• Inline Comments: mcp__github_inline_comment__create_inline_comment
• File Operations: Read,Write,Edit
• Git Operations: Bash(git:*)

Best Practices

• Be specific in your prompts
• Include expected output format
• Set clear success criteria
• Provide context about the repository
• Use inline comments for code-specific feedback