code-review

Run eight parallel agents to audit a codebase and compile a dated markdown report.

18|2|Updated Feb 6, 2026
One-click install
npx skills add https://github.com/jhd3197/CachiBot --skill code-review-jhd3197
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-review
Source: https://github.com/jhd3197/CachiBot/tree/main/.claude/skills/code-review
Command: npx skills add https://github.com/jhd3197/CachiBot --skill code-review-jhd3197

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

A coordinated, multi-perspective audit is needed to find functional bugs, security vulnerabilities, API inconsistencies, duplicated logic, test coverage gaps, state issues, and observability blindspots across a large repository; manual reviews are slow and error-prone. This Skill orchestrates eight specialized review agents to scan the codebase in parallel and produce a single, actionable audit with file paths and line numbers for each finding.

Core Features & Use Cases

  • Eight specialized agents (DRY, Security, API, Data, Frontend, Tests, State, Observability) run in parallel to provide focused, non-overlapping coverage.
  • Structured findings labeled Fix / Improve / Note with exact file paths and line numbers to enable quick triage and remediation.
  • Repository-wide scope that includes backend routes, storage, frontend source, stores and API clients, tests, and observability/logging code.
  • Automated report generation into a dated markdown file under .reviews/YYYY-MM-DD-review.md to track progress across review iterations.
  • Use Case: Before a major release, run the 8-agent sweep to surface urgent security fixes, API contract mismatches, missing tests for critical endpoints, and repeated code patterns that should be refactored into shared abstractions.

Quick Start

Run a full 8-agent sweep of the repository and output a consolidated dated report to .reviews/YYYY-MM-DD-review.md.

Frequently Asked Questions about code-review

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

FAQPage Schema
How do I run a multi-agent codebase audit for security vulnerabilities and API inconsistencies?

A multi-agent codebase audit orchestrates specialized agents to scan repository directories in parallel, identifying security vulnerabilities, API mismatches, and test coverage gaps with severity labels, file paths, and line numbers in a consolidated markdown report.

What's the best way to find duplicated logic and test coverage gaps before a major release?

Finding duplicated logic and test coverage gaps before a release requires running parallel audit agents across backend routes, frontend stores, and tests, producing structured findings labeled by severity with exact file locations for quick triage and remediation.

Can I use a parallel code review tool to scan both frontend and backend directories simultaneously?

Yes, parallel code review agents execute concurrently using exploration subagents, scanning frontend source, backend routes, storage, and tests simultaneously to surface contract mismatches, state problems, and observability faults across the entire repository.

How does an eight-agent code review sweep categorize findings for triage?

An eight-agent code review sweep categorizes findings as Fix, Improve, or Note, assigning severity labels with exact file paths and line numbers to each issue, enabling developers to quickly prioritize urgent security fixes and refactoring opportunities.

Does automated codebase auditing generate reports with file paths and line numbers for each finding?

Automated codebase auditing compiles all agent findings into a dated markdown report at .reviews/YYYY-MM-DD-review.md, containing severity-labeled issues with precise file paths and line numbers to track progress across review iterations.

What types of code issues can a multi-agent static analysis audit detect across a repository?

A multi-agent static analysis audit detects functional bugs, security vulnerabilities, API contract inconsistencies, duplicated logic patterns, test coverage gaps, state management problems, and observability blindspots across backend, frontend, and test directories.