caveman-review

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

Updated Mar 30, 2026
One-click install
npx skills add https://github.com/TrayMachi/dotfiles --skill caveman-review-traymachi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: caveman-review
Source: https://github.com/TrayMachi/dotfiles/tree/main/agents/skills/caveman-review
Command: npx skills add https://github.com/TrayMachi/dotfiles --skill caveman-review-traymachi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? PR review feedback is often buried under hedging, filler phrases, and restated diff content, making comments slow to write and slow to read. This Skill compresses every finding into a single actionable line so reviewers communicate faster without losing signal. ## Core Features & Use Cases - Terse comment format: Each finding follows L<line>: <severity> <problem>. <fix>. with exact line numbers and backticked symbol names. - Severity tagging: Uses 🔴 bug, 🟡 risk, 🔵 nit, and ❓ question prefixes to prioritize findings at a glance. - Auto-clarity mode: Automatically expands to full paragraphs for CVE-class security findings, architectural disagreements, and onboarding contexts, then resumes terse output. - Use Case: Paste a pull request diff and ask for a review; receive paste-ready one-line comments like L42: 🔴 bug: user can be null after .find(). Add guard before .email. ## Quick Start Ask the assistant to review this PR diff using caveman-review and output one-line comments with severity tags.

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 pull request?

Use a one-line format per finding: line number, severity tag, the problem, and a concrete fix. Drop phrases like "I noticed that" or "you might want to consider" and keep exact symbol names in backticks so the author can act immediately.

How to trigger the caveman-review skill?

Invoke it with /caveman-review or by saying "review this PR", "code review", or "review the diff". It auto-triggers when reviewing pull requests and outputs paste-ready comments.

What severity levels should code review comments use?

This Skill uses four levels: 🔴 bug for broken behavior, 🟡 risk for fragile code like missing null checks or swallowed errors, 🔵 nit for style issues the author can ignore, and ❓ for genuine questions.

When should code review comments be verbose instead of terse?

Terse mode should be dropped for CVE-class security findings, architectural disagreements, and onboarding contexts where the author needs the reasoning. The Skill writes a full paragraph for those cases, then resumes one-line comments.

Does caveman-review approve PRs or run linters?

No. It only outputs review comments ready to paste into the PR. It does not write code fixes, approve or request changes, or run linters; saying "stop caveman-review" reverts to verbose review style.