Add Claude Code GitHub Workflow

[allisoneer]

🤖 Installing Claude Code GitHub App

This PR adds a GitHub Actions workflow that enables Claude Code integration in our repository.

What is Claude Code?

Claude Code is an AI coding agent that can help with:

• Bug fixes and improvements
• Documentation updates
• Implementing new features
• Code reviews and suggestions
• Writing tests
• And more!

How it works

Once this PR is merged, we'll be able to interact with Claude by mentioning @claude in a pull request or issue comment.
Once the workflow is triggered, Claude will analyze the comment and surrounding context, and execute on the request in a GitHub action.

Important Notes

• This workflow won't take effect until this PR is merged
• @claude mentions won't work until after the merge is complete
• The workflow runs automatically whenever Claude is mentioned in PR or issue comments
• Claude gets access to the entire PR or issue context including files, diffs, and previous comments

Security

• Our Anthropic API key is securely stored as a GitHub Actions secret
• Only users with write access to the repository can trigger the workflow
• All Claude runs are stored in the GitHub Actions run history
• Claude's default tools are limited to reading/writing files and interacting with our repo by creating comments, branches, and commits.
• We can add more allowed tools by adding them to the workflow file like:

allowed_tools: Bash(npm install),Bash(npm run build),Bash(npm run lint),Bash(npm run test)

There's more information in the Claude Code documentation.

After merging this PR, let's try mentioning @claude in a comment on any PR to get started!

---

Important

Adds GitHub Actions workflows for Claude Code integration to automate code reviews and respond to '@claude' mentions in comments and issues.

• Workflows:
  • Adds .github/workflows/claude-code-review.yml to automate code reviews on PRs when opened or synchronized.
  • Adds .github/workflows/claude.yml to trigger Claude on issue comments, PR review comments, and issue events when '@claude' is mentioned.
• Permissions:
  • Both workflows run on ubuntu-latest and require read permissions for contents, pull-requests, and issues, and write for id-token.
• Security:
  • Uses ANTHROPIC_API_KEY stored as a GitHub Actions secret.
  • Only users with write access can trigger workflows.
• Customization:
  • Supports optional model specification, trigger phrases, and allowed tools for running commands.
  • Provides direct prompts for automated reviews and allows customization based on file types or author.

^{This description was created by for 78a0118. You can customize this summary. It will automatically update as commits are pushed.}

[claude]

Claude finished @AdjectiveAllison's task —— View job

---

Code Review for Claude Code GitHub Workflows

I've reviewed both workflow files completely. Here's my analysis following the Developer Agent approach:

Todo List:

• Read CLAUDE.md for repository-specific instructions
• Adopt Developer Agent persona (Dan Abramov approach)
• Read and analyze .github/workflows/claude-code-review.yml (complete file)
• Read and analyze .github/workflows/claude.yml (complete file)
• Evaluate security considerations
• Check for best practices and potential issues
• Provide comprehensive review feedback

---

✅ Overall Assessment: GOOD FOUNDATION

The workflows are well-structured and follow GitHub Actions best practices. However, following Dan Abramov's principle of "delete more than you add," I found several areas for improvement.

🔍 Detailed Review

.github/workflows/claude-code-review.yml (75 lines)

✅ Strengths:

• Clear trigger conditions (pull_request: [opened, synchronize])
• Appropriate permissions structure
• Good use of direct_prompt for automated reviews
• Excellent commented examples for customization
• Uses specific action version (@v4)

⚠️ Issues Found:

1. Line 32: Potentially limiting fetch depth

   fetch-depth: 1

   • This might limit Claude's ability to understand full context for complex PRs
   • Consider removing or increasing for better code analysis

2. Missing write permissions

   permissions:
     contents: read
     pull-requests: read  # Should be 'write' for Claude to comment
     issues: read

3. Redundant commented code (Lines 6-11, 15-19, 54-75)

   • 30+ lines of commented examples that could be in documentation
   • Following "delete more than you add" principle: move to docs

.github/workflows/claude.yml (59 lines)

✅ Strengths:

• Comprehensive trigger coverage
• Logical conditional structure
• Good security with API key handling
• Flexible customization options

⚠️ Issues Found:

1. Lines 15-19: Overly complex conditional logic

   if: |
     (github.event_name == 'issue_comment' && contains(github.event.comment.body, '@claude')) ||
     (github.event_name == 'pull_request_review_comment' && contains(github.event.comment.body, '@claude')) ||
     (github.event_name == 'pull_request_review' && contains(github.event.review.body, '@claude')) ||
     (github.event_name == 'issues' && (contains(github.event.issue.body, '@claude') || contains(github.event.issue.title, '@claude')))

   • Simplification opportunity: Extract common pattern contains(..., '@claude')
   • Could be more readable with a helper function or cleaner logic

2. Same permission issue as above

3. Lines 40-59: Excessive commented examples

   • 20 lines of commented configuration examples
   • Should be documentation, not code

🔒 Security Analysis

✅ Secure practices:

• API key properly stored in secrets
• Appropriate permission scoping
• No hardcoded credentials

⚠️ Security concerns:

• Missing pull-requests: write permission might cause failures
• No explicit branch protection or user restrictions
• id-token: write might be unnecessarily broad

🚀 Performance Considerations

✅ Good performance:

• fetch-depth: 1 for faster checkouts
• Efficient trigger conditions
• No unnecessary resource usage

⚠️ Potential optimizations:

• Consider caching for repeated runs
• Fetch depth might need adjustment for context

📋 Recommendations (Following Dan Abramov's "Delete More Than You Add")

1. DELETE: 50+ lines of commented examples

   • Move to documentation or README
   • Keep workflows clean and focused

2. FIX: Permissions

   permissions:
     contents: read
     pull-requests: write  # Changed from read
     issues: write         # Changed from read
     id-token: write

3. SIMPLIFY: Conditional logic in claude.yml

   • Extract common patterns
   • Make more readable

4. CONSIDER: Fetch depth

   • Test if fetch-depth: 1 limits Claude's context
   • May need fetch-depth: 0 for full history

🏗️ Integration with Agent Control Plane

Given this is a Kubernetes operator project:

• ✅ Workflows align with the project's automation needs
• ✅ Will help with code review for complex Kubernetes controllers
• ⚠️ Consider adding allowed_tools for Go/Kubernetes commands:

  allowed_tools: "Bash(make -C acp fmt vet lint test),Bash(kubectl get pods)"

📈 Test Coverage

While these are workflow files, consider:

• Testing different trigger scenarios
• Validating permissions work correctly
• Ensuring Claude can actually comment on PRs

---

🎯 Final Recommendation: APPROVE with suggested improvements

The workflows provide solid Claude Code integration. The main issues are:

1. Too much commented code (delete 50+ lines)
2. Incorrect permissions (need write access)
3. Overly complex conditionals (simplify)

These are well-architected workflows that follow GitHub Actions best practices. The foundation is solid for enabling AI-assisted code reviews and development.

---

[ellipsis-dev]

Important

Looks good to me! 👍

Reviewed everything up to 78a0118 in 50 seconds. Click for details.

• Reviewed 146 lines of code in 2 files
• Skipped 0 files when reviewing.
• Skipped posting 3 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. .github/workflows/claude-code-review.yml:36

• Draft comment:
  Consider pinning the 'anthropics/claude-code-action' version to a stable release or specific commit (instead of using '@beta') to avoid unexpected issues from beta changes.
• Reason this comment was not posted:
  Confidence changes required: 33% <= threshold 50% None

2. .github/workflows/claude.yml:34

• Draft comment:
  Consider pinning the 'anthropics/claude-code-action' version to a stable release or commit hash rather than the '@beta' tag for improved reliability.
• Reason this comment was not posted:
  Confidence changes required: 33% <= threshold 50% None

3. .github/workflows/claude.yml:15

• Draft comment:
  The complex 'if' condition relies on event payload fields. Consider adding safeguards or splitting complex conditions to ensure robustness in case expected properties are missing.
• Reason this comment was not posted:
  Confidence changes required: 33% <= threshold 50% None

Workflow ID: wflow_Y29YvmO22Qj3mvpc

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}