caveman-review

Generate terse one-line code review comments for PR diffs.

99|6|Updated Jan 11, 2026
One-click install
npx skills add https://github.com/vvedantb/eva --skill caveman-review-vvedantb
Or copy as Structured Prompt for Agentβ–Ό
Please help me install this Agent Skill.
Skill: caveman-review
Source: https://github.com/vvedantb/eva/tree/main/.agents/skills/caveman-review
Command: npx skills add https://github.com/vvedantb/eva --skill caveman-review-vvedantb

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Code review noise is reduced by producing ultra-compact, actionable one-line feedback per issue, preserving the essential signal.

Core Features & Use Cases

  • One-line findings: Each issue is summarized as a single, precise line with location and fix.
  • Structured format & severities: Supports L<line>: <problem>. <fix> format and optional severity prefixes like πŸ”΄, 🟑, or πŸ”΅ to convey importance.
  • Workflow triggers & guidance: Auto-triggers when reviewing pull requests and provides clear drop/keep guidance to stay concise.
  • Use Case: When a user asks to review a PR, the tool outputs terse, immediately actionable feedback suitable for fast iteration.

Quick Start

Ask caveman-review to review the current PR diff and generate terse, actionable comments.

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 one-line comments for a pull request diff?β–Ό

To generate concise one-line comments for a pull request diff, use an automated review tool that enforces a structured format like `L<line>: <problem>. <fix>`. This approach summarizes each issue into a single actionable line with exact line references.

Can I automate code reviews for multi-file PR diffs without generating verbose feedback?β–Ό

Yes, you can automate code reviews for multi-file PR diffs to avoid verbose feedback. The review process supports PR-triggered automation across single or multi-file diffs, producing ultra-compact, actionable feedback per issue instead of long explanations.

What is the best way to format code review comments for fast iteration?β–Ό

The best way to format code review comments for fast iteration is using a terse, one-line-per-finding structure. This format pairs exact line references with a problem statement and concise fix, optionally adding severity prefixes like πŸ”΄, 🟑, or πŸ”΅ to convey importance quickly.

How does severity prefixing work in automated code review tools?β–Ό

Severity prefixing in automated code review tools works by adding visual indicators like πŸ”΄, 🟑, or πŸ”΅ before a one-line finding. This conveys the issue's importance level immediately, helping developers prioritize fixes during fast iteration cycles.

Does this automated review format support single file diffs or is it limited to multi-file pull requests?β–Ό

This automated review format supports both single and multi-file pull request diffs. It applies the same terse, one-line-per-finding feedback structure across any PR diff size, ensuring consistent, actionable output regardless of scope.

Why does my automated code review output lack exact line references for PR findings?β–Ό

Automated code review output may lack exact line references if the tool does not enforce a strict formatting structure. Using a format that mandates `L<line>: <problem>. <fix>` ensures every finding includes precise line numbers and concise fixes.