PR Brief — AI skill for Claude Code
quick skill to facillitate green-field/big-bang reviews with AI.
How to install PR Brief
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open lucastononro/pr-brief and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What PR Brief does
quick skill to facillitate green-field/big-bang reviews with AI.
Alternatives in AI
- Field Lab — A field lab for thinking with AI 570 ★
- Maestro Companion — Quick execution for small tasks — minimal run lifecycle (start + done) with evidence recording 530 ★
- Nano Banana 2 Skill — AI image generation CLI powered by Gemini 3 Pro 228 ★
README
pr-brief — Claude Code marketplace
An interactive, browser-based PR review workflow for [Claude Code](https://docs.claude.com/en/docs/claude-code/overview). It pulls a pull request via `gh`, asks an Opus agent to group every changed file into features and author a **narrative layer** of "explain pills", and launches a local UI with GitHub-style stacked diffs, click-and-drag multi-line inline commenting, and realtime comment posting.
This repository is a **plugin marketplace**. It ships one plugin: `pr-brief`.
What reviewers get:
- A reviewer-ordered file list (migrations first, tests last) grouped into the features the PR actually contains
- Two parallel narrative layers next to each diff:
- Purple ✨ explain pills — short AI-authored callouts that connect each block to the broader change
- Red ⚠ critique pills with a severity score S1–S5 — actual review concerns, optional
suggested_changeblocks, and a "Use as comment →" button that turns a critique into an inline GitHub comment in one click
- Click-and-drag multi-line inline comments that POST to GitHub in realtime via
gh api - 💬 Inline chat: right-click any selected snippet → a streaming chat panel pops on the right that resumes the same Claude Code session that wrote the brief, so it already has full PR context. (details below)
- Runs locally; the only external call is to the Anthropic API your Claude Code session is already using
Demo
[](https://youtu.be/9d9u7UEAgLU)
Why this exists
Three reasons:
- I had to review ginormous PRs and I couldn't afford asking folks to break their PRs and stack them.
- Reviewing PRs is somewhat of finding needles in a haystack — with this I can know the right sequence to look.
- Reading code isn't always straightforward; you have to connect the dots between files mentally. This facilitates it.
Install
I
Related Skills
Px0
px0 is the IDE for humans and AI, optimized for quick, fast code reviews. It turns your browser into a zero-la
Karajan Code
Your AI writes the code; Karajan governs how it happens: TDD-first method, cross-AI review on every commit and
Aloop
Claude Code plugin for agentic agile + autonomous TDD: hook-enforced AI coding agents that plan sprints, then
TikTok Ad Compliance
Claude skill for creating TikTok ads (video, image, animated, Spark) that pass TikTok ad review and comply wit
Agent Metro
One metro map for all your projects: a plain-text TODO.md where each project is a line and each task a station
Monster Roommates
Monster Roommates: move monster tenants in so every house rule stays green. Every house has exactly one answer
Related Agents
Market Segment Reviewer
Reviews a green-bucket product idea card for market-segment soundness — is the segment crisp and big enough, r
Big Picture Reviewer
Reviews PRs for alignment with stated goals, detects "CI chasing" anti-patterns, and ensures AI agent/skill in
Deal Coach
Reviews an open deal and coaches the next move: qualification gaps, stakeholders, risks, objections and a conc