What problem does it solve? Review comments on pull requests are often written in the reviewer's own vocabulary and assumptions, making it hard to judge what is actually being flagged and whether to act on it. This Skill retrieves unresolved review comments from a GitHub PR and breaks each one down into the intent, the current code, a proposed change, its impact, and a recommendation, so you can decide how to respond without guesswork. ## Core Features & Use Cases - Comment retrieval and filtering: Collects inline comments, review bodies, and PR conversation comments via the GitHub CLI, excluding command-only comments and already-resolved threads, with optional filtering by comment ID or criteria. - Human vs bot classification: Distinguishes commenters via the GitHub user type so follow-up work routes to the correct resolution workflow (/resolve-human-reviews or /resolve-ai-reviews). - Explanation with learning takeaways: For each comment, explains the underlying principle, checks whether the same issue appears elsewhere in the diff, and gives cues for spotting it yourself next time. - Use Case: You open a PR and find eight review comments from teammates and bots. Run this Skill to get a summary table plus detailed explanations of the non-trivial items, each with a recommendation, so you can decide what to fix before handing off to a resolution skill. ## Quick Start Ask the assistant to explain the review comments on the current pull request, optionally passing a PR number or a specific comment ID to narrow the scope.
