copilot-feedback-resolver

Copilot-feedback-resolver scans your PR for Copilot-generated review threads, applies code changes to address the feedback, and resolves threads using GitHub's GraphQL API. It distinguishes Copilot threads from human reviews, ensuring only automated feedback is resolved while preserving human input.

Yes. The Skill uses GraphQL mutations like resolveReviewThread and addPullRequestReviewThreadReply to reply to Copilot feedback and mark threads resolved. It handles inline mutation values and requires the gh CLI for authentication and PR context.

The Skill updates your copilot-instructions.md file to document incorrect feedback, preventing Copilot from repeating the same suggestions in future PRs. This builds institutional knowledge and reduces recurring false positives over time.

By automatically detecting Copilot threads, applying fixes, and closing resolved feedback, the Skill eliminates the manual toil of reading each comment and clicking resolve. Human reviews remain untouched, keeping your PR workflow focused on substantive feedback.

Yes. The Skill integrates into PR workflows where Copilot has already left feedback. It pushes code changes and performs explicit post-resolution verification, making it suitable for automated review pipelines that run after Copilot analysis.

You need GitHub CLI (gh) installed, repository access with PR permissions, and a GitHub token for GraphQL authentication. The Skill handles thread detection and resolution; no additional dependencies or components are required.