What problem does it solve? Turning a GitHub issue into a reviewed pull request involves many error-prone steps: assessing whether the issue is actionable, clarifying ambiguities, choosing an implementation approach, writing tests, matching CI locally, and documenting decisions. This Skill automates that entire workflow while keeping a human in control at two approval gates. ## Core Features & Use Cases - Readiness assessment and clarification: Scores the issue against a six-dimension rubric (goal, acceptance, scope, locus, testability, compatibility), asks batched clarifying questions, and posts the consolidated Q&A to the issue. - Approach selection with optional racing: Implements straightforward changes directly, or races 2-3 competing implementations in isolated git worktrees judged against a pre-written rubric for complex work. - Tests, CI mirror, and guarded delivery: Writes tests matching the repo's existing conventions, replays the GitHub Actions workflow commands locally, then pushes and opens a PR only after explicit approval. It never merges and never commits to main. - Resumable state tracking: Persists per-issue state files so an interrupted run can resume from the exact phase it reached without double-posting comments. - Use Case: Invoke with an issue number to have the Skill branch, implement, test, mirror CI, and open a pull request that closes the issue, posting exactly four structured comments along the way. ## Quick Start Ask the assistant to implement GitHub issue number 42 end-to-end on a new branch and open a pull request for review.
