What problem does it solve? Managing pull requests involves repetitive manual steps: creating branches, writing conventional commits, opening PRs, watching CI, diagnosing failures, and merging. This Skill guides an AI agent through the entire GitHub PR lifecycle with ready-to-run commands, working both with the gh CLI and with plain git plus curl against the GitHub REST and GraphQL APIs. ## Core Features & Use Cases - Full PR Lifecycle Coverage: Branch naming conventions, Conventional Commits, PR creation (including drafts, reviewers, and labels), CI status polling, and squash/rebase/merge with branch cleanup. - Dual Auth Modes: Detects whether gh is authenticated and falls back to git + curl with a GITHUB_TOKEN, including extracting owner/repo from the git remote and enabling auto-merge via GraphQL. - CI Auto-Fix Loop: A structured diagnose-fix-push-verify loop (up to 3 attempts) backed by a troubleshooting reference covering test, lint, type, build, permission, timeout, and Docker failures. - Use Case: Ask the agent to implement a fix on a new branch, open a PR that closes an issue, watch the checks, automatically diagnose and fix a failing lint job, then squash-merge and delete the branch. ## Quick Start Ask the agent to create a branch, commit your current changes with a conventional commit message, open a pull request, monitor CI, and merge it when the checks pass.
