caveman-review

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

Updated Aug 6, 2026
One-click install
npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill caveman-review-ferrarifankid04
Or copy as Structured Prompt for Agentβ–Ό
Please help me install this Agent Skill.
Skill: caveman-review
Source: https://github.com/ferrarifankid04/ai-skills-public/tree/main/antigravity/caveman/skills/caveman-review
Command: npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill caveman-review-ferrarifankid04

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Code review feedback is often buried in verbose, hedged prose that obscures the actionable point. This Skill compresses PR review comments into single lines containing the exact location, the problem, and a concrete fix, so authors can act on feedback immediately. ## Core Features & Use Cases - Terse Comment Format: Outputs findings as L<line>: <severity> <problem>. <fix>. with severity emoji (πŸ”΄ bug, 🟑 risk, πŸ”΅ nit, ❓ question) and backticked symbol names. - Noise Filtering: Drops throat-clearing, hedging, per-comment praise, and restatements of what the diff already shows. - Auto-Clarity Mode: Switches to full explanatory paragraphs for CVE-class security findings, architectural disagreements, and onboarding contexts, then resumes terse output. - Use Case: Paste a pull request diff and receive ready-to-post review comments like L42: πŸ”΄ bug: user can be null after .find(). Add guard before .email. ## Quick Start Ask the AI to review this pull request diff and return one-line caveman-style comments with severity, problem, and fix for each finding.

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?β–Ό

Use the format L<line>: <severity> <problem>. <fix>. with one line per finding. Prefix with a severity emoji (πŸ”΄ bug, 🟑 risk, πŸ”΅ nit, ❓ question), keep exact line numbers and backticked symbol names, and state a concrete fix instead of vague suggestions.

How to review a PR diff without writing long explanations?β–Ό

Invoke the skill with "review this PR" or "/caveman-review" and paste the diff. It outputs one-line findings with location, problem, and fix, dropping hedging phrases and restatements of what the code already shows.

When should code review comments be verbose instead of terse?β–Ό

Terse mode should be dropped for CVE-class security findings, architectural disagreements, and onboarding contexts where the author needs the reasoning. In those cases write a full explanatory paragraph, then resume one-line comments for the remaining findings.

Does this code review skill approve or request changes on pull requests?β–Ό

No, it only generates review comment text ready to paste into the PR. It does not approve, request changes, write code fixes, or run linters; those actions remain with the human reviewer.

What severity levels should I use in code review feedback?β–Ό

Use four levels: πŸ”΄ bug for broken behavior that will cause incidents, 🟑 risk for fragile code like missing null checks or swallowed errors, πŸ”΅ nit for style issues the author can ignore, and ❓ question for genuine inquiries rather than disguised suggestions.