caveman-review

Generate one-line code review comments with location, problem, and fix.

4|2|Updated May 18, 2022

Search Tags:#github #diff #code-review #pull-request #dev-tools #concise-feedback #review-format

Ready to run

npx skills add https://github.com/pjmagee/starwars-data --skill caveman-review-pjmagee

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: caveman-review
Source: https://github.com/pjmagee/starwars-data/tree/main/.agents/skills/caveman-review
Command: npx skills add https://github.com/pjmagee/starwars-data --skill caveman-review-pjmagee

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill reduces noisy, verbose PR feedback into ultra-compressed, actionable one-line comments that surface only location, problem, and a concrete fix so reviewers and authors iterate faster.

Core Features & Use Cases

Ultra-compressed, one-line review comments formatted as location, problem, fix.
Optional severity prefixes for bugs, risks, nits, and questions plus a rule to expand into full paragraphs for security findings or onboarding contexts.
Works for single-file diffs, multi-file pull requests, conversational triggers like review commands, and fast reviewer workflows that prioritize signal over commentary.

Quick Start

Review this pull request and return one-line caveman-review comments for each finding using location, problem, fix.

Frequently Asked Questions about caveman-review

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

FAQPage Schema

How do I write concise code review comments for a GitHub pull request diff?▼

Concise code review comments for a GitHub pull request diff are generated as ultra-compressed one-line findings. Each comment contains the location, the problem, and a concrete fix, preserving exact symbol names and line numbers to help reviewers and authors iterate faster.

Can I get one-line PR feedback for multi-file diffs instead of verbose paragraphs?▼

Yes, one-line PR feedback works for both single-file and multi-file pull request diffs. It automatically compresses findings unless a security issue or onboarding context is detected, where it expands into full paragraphs to provide necessary detail.

What is the best way to format code review findings to surface risks and nits?▼

The best way to format code review findings is using optional severity prefixes for bugs, risks, nits, and questions. This high-signal format prioritizes actionable feedback by pairing each severity tag with the exact location, problem description, and a targeted fix.

Does this code review approach work with conversational triggers like slash commands?▼

This code review approach works directly with conversational triggers such as review commands or slash commands. Applying it to a PR diff or typing a request like review this PR processes the changes and emits terse, actionable one-line findings.

When should I avoid using ultra-compressed code review formats?▼

You should avoid using ultra-compressed code review formats when dealing with security findings or onboarding contexts. In these specific situations, the format expands into full paragraphs to ensure critical information and context are not lost during the review process.