review-pr

Analyze PR size and risk to assign a review tier and dispatch prompts.

97|4|Updated Mar 25, 2026
One-click install
npx skills add https://github.com/go-to-k/cdkd --skill review-pr-go-to-k
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-pr
Source: https://github.com/go-to-k/cdkd/tree/main/.claude/skills/review-pr
Command: npx skills add https://github.com/go-to-k/cdkd --skill review-pr-go-to-k

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automatically determines the appropriate level of PR review rigor based on size and risk, and provides a concrete plan for reviewers.

Core Features & Use Cases

  • Predicts the required review tier (inline, 1-reviewer, or 3-axis) based on PR size and file changes.
  • Applies bias factors to adjust the tier when surface risk (security, tests, infra) or project context warrants it.
  • Outputs ready-to-use Agent dispatch prompts for the orchestrator when reviewers are warranted.
  • Useful for fast feedback loops in large teams, or when balancing review load across multiple PRs.

Quick Start

Run /review-pr <PR-number> to get a recommended review tier and ready-to-dispatch prompts.

Frequently Asked Questions about review-pr

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

FAQPage Schema
How do I determine the right PR review depth for a pull request?

PR review depth is determined by analyzing the pull request's size in lines of code, changed files, and baseline risk. Bias factors for surface areas like security, tests, and infrastructure adjust the assessment to produce a final review tier.

What's the best way to automate pull request review triage for large teams?

Automating pull request review triage applies defined heuristics to PR stats read via gh, predicting a required review tier. It outputs ready-to-use dispatch prompts to balance review load across multiple PRs for large teams.

How does heuristics-based PR review level prediction work?

Heuristics-based PR review prediction evaluates PR size and file changes against baseline risk, applying bias factors for security, tests, and infra surface areas. It outputs a structured plan with a final tier of inline, 1-reviewer, or 3-axis.

Can I balance PR review load automatically based on PR size and risk?

Yes, you can balance PR review load by running the review-pr command on a PR number. It evaluates size and risk to recommend an appropriate review tier and provides ready-to-dispatch prompts for orchestrating reviewers.

What are the limitations of using automated heuristics for PR review tiering?

Automated heuristics for PR review tiering rely on quantifiable metrics like LOC and changed files read via gh. They may not capture nuanced contextual risks beyond defined surface areas like security, tests, and infrastructure without manual bias adjustments.

Does the review-pr Skill require the GitHub CLI to analyze pull requests?

Yes, the review-pr Skill requires the GitHub CLI (gh) to read PR stats. It uses these statistics to apply heuristics and output a structured review plan with appropriate tiers and dispatch prompts.