caveman-review

Generate one-line code review comments with severity and fixes.

Updated Aug 27, 2026

Search Tags:#static analysis #code review #pull requests #comment generation #severity labels #diff feedback

Ready to run

npx skills add https://github.com/ebarroso12/personal-skills --skill caveman-review-ebarroso12

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: caveman-review
Source: https://github.com/ebarroso12/personal-skills/tree/main/caveman-review
Command: npx skills add https://github.com/ebarroso12/personal-skills --skill caveman-review-ebarroso12

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill cuts noisy, verbose feedback down to compact, paste-ready review comments that still preserve the actionable signal for faster, clearer code reviews.

Core Features & Use Cases

Ultra-compressed review comments: Produces one line per finding in the format L<line>: <severity> <problem>. <fix>. for quick scanning.
Action-oriented severity labels: Uses severity emojis (🔴 bug, 🟡 risk, 🔵 nit, ❓ q) to communicate impact at a glance.
Reduced ambiguity and repetition: Drops throat-clearing, hedging, and restating what the diff already shows, while keeping exact line numbers and concrete fixes.
Auto-clarity mode for important contexts: Switches to fuller explanations for security (CVE-class), architectural disagreements, and onboarding contexts where the “why” matters.

Quick Start

Ask your AI to review the diff with caveman-review by saying: review the diff for my pull request.

Frequently Asked Questions about caveman-review

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

FAQPage Schema

How do I generate concise code review comments for a pull request?▼

To generate concise code review comments, format findings as one-line, location-based feedback with severity, problem, and fix. This compresses verbose diff feedback into paste-ready text like L<line>: <severity> <problem>. <fix>.

What is the best way to add severity labels to static analysis feedback?▼

Adding severity labels to static analysis feedback uses emojis like 🔴 bug, 🟡 risk, 🔵 nit, and ❓ q. This communicates impact at a glance while keeping exact line numbers and concrete fixes.

Can I use automated comment generation for multi-file pull request diffs?▼

Yes, automated comment generation applies to single- and multi-file pull request diffs. It enforces strict formatting using <file>:L<line>: to produce actionable, paste-ready comments across multiple files.

When do I need fuller explanations instead of ultra-compressed diff feedback?▼

Fuller explanations are needed for security CVE-class issues, architectural disagreements, and onboarding contexts. Auto-clarity mode switches from one-line diff feedback to detailed explanations when the why matters.

Does ultra-compressed code review feedback drop exact line numbers and symbols?▼

No, ultra-compressed code review feedback keeps exact line numbers and symbols. It minimizes hedging, throat-clearing, and redundant explanations while preserving the actionable signal for faster scanning.