pr-fix-guide

Fixes GitHub PR review comments through root cause analysis and direct code edits.

1|1|Updated Dec 18, 2025
One-click install
npx skills add https://github.com/nexuslabs-ai/nexus --skill pr-fix-guide-nexuslabs-ai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: pr-fix-guide
Source: https://github.com/nexuslabs-ai/nexus/tree/main/.agents/skills/pr-fix-guide
Command: npx skills add https://github.com/nexuslabs-ai/nexus --skill pr-fix-guide-nexuslabs-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Addressing PR review feedback often involves slow back-and-forth between analysis and implementation, or blindly applying reviewer suggestions that treat symptoms instead of root causes. This Skill combines root cause analysis and fixing into a single autonomous flow that resolves every review comment directly. ## Core Features & Use Cases - Review Context Gathering: Fetches PR details, review threads, inline comments, and diffs via the GitHub CLI (gh pr view, gh api), then categorizes issues as blocking or minor. - Root Cause Analysis: Reads surrounding code for each flagged issue, treats reviewer suggestions as input rather than instruction, and refuses to defer findings into a follow-up bucket. - Autonomous Fixing with Verification: Emits a one-line-per-fix plan table, applies all fixes without per-fix confirmation, then verifies with pnpm typecheck and pnpm lint. - Use Case: A reviewer leaves eight comments on your pull request, two marked REQUEST_CHANGES. Invoke this Skill with the PR number and it analyzes each comment, fixes blocking issues first, and reports a compact table of what changed with clean typecheck and lint results. ## Quick Start Ask the agent to fix all review comments on pull request number 123 using the pr-fix-guide workflow.

Frequently Asked Questions about pr-fix-guide

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

FAQPage Schema
How do I fix all PR review comments automatically?▼

Invoke the Skill with a PR number and it fetches all reviews and inline comments via the GitHub CLI, categorizes them as blocking or minor, then fixes each one autonomously. It finishes by running typecheck and lint to verify the changes.

How to address GitHub review feedback without blindly applying suggestions?▼

The Skill treats reviewer suggestions as input, not instruction. For each comment it reads the surrounding code, identifies whether the flagged line is the root cause or a symptom, and fixes the underlying cause with the smallest correct change.

Does this Skill require the GitHub CLI to be installed?▼

Yes, it relies on the gh CLI for fetching PR details, reviews, comments, and diffs using commands like gh pr view and gh api. The repository must also be a GitHub-hosted project for these API calls to resolve.

Can I defer non-blocking review comments to a follow-up PR?▼

No, the Skill explicitly forbids deferral. Reviewer framing like 'not blocking' or 'follow-up' does not exempt a finding; everything is fixed in the current PR unless it maps to a cited, existing issue number.

When does the Skill pause for user input during fixes?▼

It pauses only when a fix is genuinely ambiguous, requires an architectural decision beyond the comment's scope, or conflicts with another pending fix. The initial plan table is a preview, not an approval gate.