pr-reviewer

Run isolated git worktree PR reviews against Bedrock models and emit structured JSON reports.

158|12|Updated Feb 11, 2026
One-click install
npx skills add https://github.com/nteract/nteract --skill pr-reviewer-nteract
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pr-reviewer
Source: https://github.com/nteract/nteract/tree/main/.agents/skills/pr-reviewer
Command: npx skills add https://github.com/nteract/nteract --skill pr-reviewer-nteract

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates repository-local PR reviews by running an isolated worktree against Bedrock models and generating structured, machine-readable feedback.

Core Features & Use Cases

  • Isolated PR review environment: creates a disposable git worktree for safe analysis.
  • Bedrock/opencode-backed evaluation: leverages AI-assisted review to assess diffs and implementations.
  • Output: emits a JSON report with verdict, terminal_reason, findings, model, session_id, cost, and diff metadata for automation and traceability.

Quick Start

Run uv run pr-review on a real PR URL to generate a structured JSON report in the .context/reviews directory

Frequently Asked Questions about pr-reviewer

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

FAQPage Schema
How do I automate GitHub PR reviews using AWS Bedrock models?

You can automate GitHub PR reviews using AWS Bedrock by running an isolated git worktree that evaluates branch diffs and produces a structured JSON report containing the verdict, findings, and associated cost metadata.

What is a git worktree and why is it used for PR review automation?

A git worktree creates a disposable, isolated local copy of your repository. Using a worktree for PR review automation ensures safe, deterministic analysis of branch diffs without modifying your primary working directory.

How do I generate structured JSON reports from pull request diffs?

To generate structured JSON reports from pull request diffs, execute the review workflow with opencode against a specified PR URL. The process emits a machine-readable file containing the verdict, terminal reason, findings, model, and session ID.

Can I use an external AI model as a second reviewer for branch diffs?

Yes, you can use an external AI model as a second reviewer for branch diffs. The workflow runs an isolated worktree against Bedrock models to provide deterministic, AI-assisted feedback for your review loops.

Does automated PR review work with existing GitHub automation pipelines?

Automated PR review integrates with GitHub automation pipelines by outputting machine-readable JSON reports to a designated directory. This structured format allows downstream systems to parse the verdict, findings, and cost metadata automatically.