caveman-review

Generate one-line PR review comments with location and fix.

Updated Jan 24, 2026

Search Tags:#formatting #diff #code-review #pull-request #lineno #comment-automation #severities

Ready to run

npx skills add https://github.com/Ikki-Group/erp --skill caveman-review-ikki-group

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: caveman-review
Source: https://github.com/Ikki-Group/erp/tree/main/.agents/skills/caveman-review
Command: npx skills add https://github.com/Ikki-Group/erp --skill caveman-review-ikki-group

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Reduces PR review noise by producing ultra-compact, actionable one-line comments that preserve the actionable signal.

Core Features & Use Cases

One-line comments per finding with exact location and identifiers.
Supports severity prefixes and deterministic formatting for easy copy-paste into PRs.
Use cases include reviewing diffs in single-file or multi-file changes and when the reviewer asks for a caveman review.

Quick Start

Review a pull request diff and generate one-line, location-specific comments with problem and fix.

Frequently Asked Questions about caveman-review

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

FAQPage Schema

How do I generate compact one-line comments for a pull request diff?▼

To generate compact one-line comments for a pull request diff, invoke the /caveman-review command during your PR review. It produces ultra-compact, actionable comments formatted as L<line>: <problem>. <fix>.

Can I review multi-file diffs without losing track of specific line numbers?▼

Yes, you can review multi-file diffs without losing track of specific line numbers. The review process generates deterministic, location-specific comments that preserve exact identifiers and line numbers across all changed files.

Does the code review output include severity prefixes for identified issues?▼

Yes, the code review output includes optional severity prefixes for identified issues. This allows you to categorize the importance of findings while maintaining an ultra-compact, single-line format for easy copy-pasting.

What is the best way to reduce noise during automated PR code reviews?▼

The best way to reduce noise during automated PR code reviews is to use a formatting method that produces ultra-compact, actionable one-line comments. This preserves the actionable signal by filtering out unnecessary verbose explanations.

How does comment-automation format findings for easy copy-paste into pull requests?▼

Comment-automation formats findings for easy copy-paste into pull requests by using a deterministic structure of L<line>: <problem>. <fix>. This precise format preserves exact code identifiers and enables immediate pasting into the PR interface.