code-review

PackAI & models

code-review is a plugin that has an AI agent automatically review pull requests and flag likely issues. It runs multiple specialized agents on each change and uses confidence-based scoring to filter out false positives, so the reviewer sees a shorter list of findings worth acting on.

Unavailable. Delivery for this kind is on the roadmap — not serving yet.

Have an AI agent setup that supports installing plugins.

What your AI can do with it

  • Automatically checks pull requests for likely issues
  • Runs multiple specialized review agents on each change
  • Uses confidence-based scoring to filter false positives
  • Flags findings so a human reviewer can act on them

Getting started

  1. Have an AI agent setup that supports installing plugins.
  2. Add the code-review plugin to that setup.
  3. Open or update a pull request and let the agent run its review.
  4. Read the flagged issues and their confidence before acting on them.

Signals

GitHub stars
37k
Forks
4k
Last commit
Sep 2026

Questions

What is this code review Claude plugin?
It is a plugin that lets an AI agent automatically review pull requests and flag likely issues, using multiple specialized agents with confidence-based scoring to filter false positives.
How does it avoid false positives?
Findings from its specialized agents go through confidence-based scoring, and low-confidence results are filtered out before they reach the reviewer.
Does it replace human review?
No. It flags likely issues on pull requests; a person still reads the findings and decides what to change.
Advanced
Item type
plugin
Key
anthropics-claude-plugins-official-code-review
Source
github.com/anthropics/claude-plugins-official