PR Workflow Plugin
Automate your pull request workflow from commit to review.
Installation
# Install the plugin
claude plugin install pr-workflow@personio-claude-code-marketplace
What does this plugin do?
This plugin handles the complete pull request workflow - from shipping new code to addressing reviewer feedback.
/ship Command
Run /ship to automatically:
- Run linting and attempt auto-fixes
- Create meaningful commit messages by analyzing your actual changes
- Create feature branches with JIRA ticket integration
- Push to GitHub and create a draft PR
- Suggest reviewers based on who wrote the code you're changing
What you need:
- Git repository
- GitHub CLI (
gh) installed and authenticated:brew install gh && gh auth login - Optional: JIRA/Atlassian MCP server for automatic branch naming from tickets
Quick Start:
# When you have changes ready to ship
/ship
That's it. Claude will guide you through the rest.
Example:
$ /ship
# Claude detects you're on main and asks about creating a branch
# You provide JIRA ticket: TEAM-1234
# Claude fetches ticket: "Add webhook retry logic"
# Creates branch: TEAM-1234/add-webhook-retry-logic
# Runs linting, analyzes your changes
# Creates commit: "[TEAM-1234] Add webhook retry mechanism with exponential backoff"
# Pushes and creates draft PR
# Suggests reviewers who wrote the code you modified
Why would you want this?
- You want to ship code faster with less manual work
- You want better commit messages without the effort
- You want to automatically find the right reviewers
- You use JIRA and want branch names synced to tickets
/address-pr-comments Command
Run /address-pr-comments to systematically address reviewer feedback on pull requests:
- Fetch PR details and all reviewer comments automatically
- Analyze changes and categorize feedback by type (bugs, improvements, style)
- Create a comprehensive fix plan for your approval
- Implement all code changes to address comments
- Optionally run linting/validation before committing
- Commit and push changes with context-aware commit messages
What you need:
- Git repository with a PR branch checked out
- GitHub CLI (
gh) installed and authenticated:brew install gh && gh auth login - Write access to the repository
Quick Start:
# Address comments on a specific PR
/address-pr-comments https://github.com/owner/repo/pull/123
# Or use shorthand
/address-pr-comments #123
# Or let Claude infer the PR from your current branch
/address-pr-comments
Claude will guide you through the rest.
Example:
$ /address-pr-comments #456
# Claude fetches PR details and comments
# Analyzes all reviewer feedback
# Creates a fix plan showing what will be changed
# You approve the plan
# Claude implements all fixes
# Asks if you want to run linting (optional)
# Generates commit message:
# [TEAM-456] Address PR review comments: fix null safety and error handling
#
# - Fixed: Null pointer risk in getUserById method
# - Added: Proper error handling for missing users
# - Improved: Error messages for better debugging
# Commits and pushes to your PR branch
# You're done!
Why would you want this?
- You want to address PR comments quickly and systematically
- You want to ensure no reviewer feedback is missed
- You want context-aware commit messages that reference the fixes
- You want to validate changes before pushing (optional linting)
- You want a complete workflow from feedback to pushed fix
pr-hints-for-reviewer Skill
Generates (or updates) a Reviewer hints section at the end of your PR description, answering a fixed set of questions about the diff to help reviewers focus on what matters.
What you need:
- GitHub CLI (
gh) installed and authenticated - A PR already open (or just created in the current session)
Quick Start:
# After opening a PR
"Generate hints for the reviewer"
# Or with an explicit PR link
"Generate hints for the reviewer of https://github.com/org/repo/pull/123"
# To refresh existing hints
"Update hints for the reviewer"
The hints section covers:
- Whether all changes are relevant to the PR's stated purpose
- Which files or hunks deserve the most careful attention
- Any naming or convention violations worth flagging
The section is wrapped in HTML comment anchors (<!-- reviewer-hints-start --> / <!-- reviewer-hints-end -->), so re-running the skill replaces the previous hints rather than appending a duplicate.
Why would you want this?
- You want to save reviewers time by pointing them at the most important changes
- You want a consistent, reproducible hints format across all your PRs
- You want a second opinion on whether your PR is focused and convention-compliant
suggest-reviewers Skill
When you create a PR, Claude can analyze git blame on your changed files to suggest the best reviewers - people who actually worked on that code.
code-reviewer Agent
Claude can automatically invoke this expert agent when you need comprehensive code review. The agent specializes in:
- Security analysis: OWASP Top 10 vulnerabilities, authentication, cryptography
- Performance optimization: Database queries, memory leaks, caching strategies
- Production reliability: Configuration security, infrastructure review, error handling
- Code quality: Clean Code principles, SOLID patterns, maintainability
- Modern tooling: Integration with SonarQube, CodeQL, Semgrep, Snyk
What you need: Just your code that needs reviewing.
Quick Start:
# Just ask Claude to review code
"Review this authentication implementation for security issues"
"Check this database query for performance problems"
"Analyze this API endpoint for production readiness"
Claude will automatically invoke the code-reviewer agent to provide expert analysis.
Example:
$ "Review the UserService.kt file for security vulnerabilities"
# Claude invokes the code-reviewer agent which:
# - Analyzes authentication and authorization logic
# - Checks for SQL injection vulnerabilities
# - Reviews input validation and sanitization
# - Assesses error handling and logging
# - Provides specific, actionable feedback with code examples
Why would you want this?
- You want expert-level code review before production deployment
- You need to identify security vulnerabilities early
- You want to catch performance issues before they hit production
- You need help maintaining code quality standards