reviews

Automate GitHub pull request review feedback triage and code fixes.

Updated Feb 2, 2026
One-click install
npx skills add https://github.com/dupe-com/skills --skill reviews-dupe-com
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: reviews
Source: https://github.com/dupe-com/skills/tree/main/skills/reviews
Command: npx skills add https://github.com/dupe-com/skills --skill reviews-dupe-com

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Manually addressing GitHub pull request review feedback is tedious and error-prone, especially when managing dozens of inline comments, conversation threads, and follow-up actions across large code changes. This Skill automates the entire end-to-end workflow to eliminate repetitive work, reduce human error, and ensure no review feedback falls through the cracks.

Core Features & Use Cases

  • Full feedback aggregation: Fetches all top-level reviews, inline line-by-line comments, and conversation-tab comments in parallel via GitHub REST and GraphQL APIs.
  • Intelligent triage: Automatically categorizes feedback as actionable, ambiguous, questions, or out-of-scope, skipping resolved, outdated, or irrelevant comments.
  • End-to-end handling: Applies code fixes for actionable items, posts helpful replies to clarify changes for reviewers, resolves addressed threads, and generates a structured summary comment for the PR.
  • Use case example: When a reviewer leaves 15 inline comments requesting code changes, clarification questions, and a few out-of-scope suggestions across a feature PR, this Skill can apply all valid fixes, reply to questions, propose follow-up issues for out-of-scope items, and post a clear summary of all actions taken, all without pushing changes for human review.

Quick Start

Use the reviews skill to address all review feedback on your current branch's open pull request, apply code fixes for actionable comments, and post a summary of all changes made.

Frequently Asked Questions about reviews

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate GitHub pull request review feedback?

Automate GitHub pull request review feedback by using an AI agent to fetch inline comments, triage actionable items, apply code fixes, and post resolution summaries. The process aggregates all review threads via GitHub APIs to ensure no feedback is missed.

Can I automatically apply code changes from PR review comments?

Yes, you can automatically apply code changes from PR review comments. The workflow categorizes feedback as actionable, ambiguous, or out-of-scope, applies fixes for actionable items, and escalates ambiguous feedback to the user for confirmation without pushing changes.

Do I need the gh CLI installed to automate PR feedback workflows?

Yes, you need the gh CLI installed and authenticated with write access to the target repository to automate PR feedback workflows. The process relies on the gh CLI to interact with GitHub REST and GraphQL APIs for fetching comments and posting replies.

What is the best way to triage ambiguous GitHub PR comments?

The best way to triage ambiguous GitHub PR comments is to automatically categorize them as actionable, ambiguous, questions, or out-of-scope. Ambiguous items are escalated to the user for confirmation, while actionable feedback receives automated code fixes and replies.

How does automated PR feedback handling deal with out-of-scope suggestions?

Automated PR feedback handling deals with out-of-scope suggestions by categorizing them separately from actionable items. It skips irrelevant comments and can propose follow-up issues for out-of-scope items, posting a clear summary of all actions taken on the pull request.

What are the limitations of automating GitHub code review responses?

Limitations of automating GitHub code review responses include strict safety guardrails where the agent never pushes changes directly and never modifies other users' comments. It requires human review of applied fixes, and ambiguous feedback must be escalated to the user.