What problem does it solve? Iterating on a draft pull request through repeated code reviews, bug audits, and bot feedback is slow and manual. This Skill automates the entire loop: it runs parallel review lenses, applies one-commit fixes, and gates readiness on Bugbot, Copilot, and Codex confirmations until the PR is ready. ## Core Features & Use Cases - Autonomous convergence loop: Runs three parallel review lenses (code-review, bug-audit, self-review) per round and lands fixes in single commits via a workflow script or portable in-session driver. - Terminal reviewer gates: Treats Cursor Bugbot, GitHub Copilot, and Codex reviews as confirmation gates before marking the PR ready, with quota and outage handling for each. - Multi-PR fan-out and reporting: Converges several PRs in parallel worktrees and publishes a closing HTML convergence report with an idempotent PR comment. - Use Case: A developer opens a draft PR and says "/autoconverge"; the Skill enters a worktree, runs review-fix rounds, verifies Copilot findings with executed checks, and marks the PR ready once all reviewers are clean on the same HEAD. ## Quick Start Ask the AI to run /autoconverge on your draft pull request URL to drive it to ready autonomously.