What problem does it solve? Handling GitHub pull request review feedback is tedious: you must find unresolved threads, decide which comments need code changes, fix them one by one, and reply to each reviewer. This Skill automates that entire loop while keeping you in control of the response plan. ## Core Features & Use Cases - Unresolved Thread Extraction: Uses the GitHub GraphQL API to fetch only unresolved review threads, with pagination support for PRs with over 100 threads. - Approval-Gated Workflow: Presents a table of each comment with a verdict (fix needed or not applicable) and waits for your approval before touching any code. - One-by-One Fix and Reply: For each comment, it fixes the code, runs fmt/clippy/tests, commits with Conventional Commits, pushes, and replies in the thread with the commit hash, matching the reviewer's language. - Knowledge Feedback Loop: When AI review bots (CodeRabbit, Devin, Copilot, Claude) raise valid points, it records the lesson into relevant skills or docs after your confirmation. - Use Case: You open a PR on an open-source Rust project and receive 12 review comments overnight. Run this Skill to triage all unresolved threads, approve a response plan, and have each item fixed, committed, and answered individually. ## Quick Start Ask the assistant to handle review comments on a pull request, for example by saying: respond to the review comments on PR 42.