skills.rest
joeledwardson
joeledwardsonCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

review-code

Reviews changed code through multi-pass architectural, per-file, and holistic diff analysis using Gemini models.

1|1|Updated Nov 25, 2024
Search Tags:#static analysis#code review#pull request#design review#gemini#git diff#llm cli
One-click install
npx skills add https://github.com/joeledwardson/dev-setup --skill review-code-joeledwardson
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: review-code
Source: https://github.com/joeledwardson/dev-setup/tree/main/configs/claude/skills/review-code
Command: npx skills add https://github.com/joeledwardson/dev-setup --skill review-code-joeledwardson

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Self-reviewing code with full project context normalizes away unclear responsibilities, hidden assumptions, and reinvented wheels. This Skill runs a cold-read review of all files changed since main, catching design drift and logic issues that the original author overlooks. ## Core Features & Use Cases - Four-pass review pipeline: architectural sweep of the full codebase, per-file design review, per-unit logic review, and a holistic diff review. - Cold-read framing: isolated single-file calls via llm -a ensure the reviewer model has no prior context, surfacing weak contracts, unclear naming, and unnecessary complexity. - Structured report output: writes flagged findings to docs/appendix/reviews/YYYY-MM-DD-<basename>.md and appends to an index, without modifying any code. - Use Case: After finishing a feature branch, run the review to get a junior-developer comprehension check on every changed Go or Python file before opening a pull request. ## Quick Start Run the review-code skill on all files changed since main to generate a flagged findings report.

Frequently Asked Questions about review-code

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

FAQPage Schema
How do I review code changes before a pull request with AI?▼

Run the review-code skill with no arguments to analyze all files changed since main, or pass a specific file path. It runs four passes—architectural sweep, per-file design review, per-unit logic review, and holistic diff review—and writes flagged findings to a dated Markdown report.

What tools are needed for multi-pass AI code review?▼

This workflow uses the gemini CLI for the full-codebase architectural pass and the llm CLI with gemini-2.5-flash for isolated per-file and diff passes. Git is required to compute changed files and diffs against main.

Does this code review modify my source files?▼

No, the review is read-only. It only writes a report to docs/appendix/reviews/ and appends a row to the index file, leaving the report unstaged and never touching source code.

Which file types does the code review support?▼

It supports Go, Python, and any source file with recognizable structure. Generated files such as *.pb.go, *_generated.go, and docs/grading/api/ are explicitly skipped.

When does the review write a report versus stay silent?▼

A report is written only when any answer across all passes contains FLAG or YES. If everything is clear, it prints 'review complete — no flags' and writes nothing.

Related Skills

Explore More Software Engineering Skills→
SeventeenthEarthSeventeenthEarth

17th-pr-summary

Generate standardized PR titles and summaries from git diffs and commit history.

Community
Basic
swarm-ai-researchswarm-ai-research

PR Review

Generate severity-tagged findings, inline comments, and verdicts for pull requests.

Official
Advanced
zajcazajca
2

go-linting

Configure golangci-lint with recommended Go linters for static analysis.

Community
Basic
smith-hornsmith-horn
75

Pair Programming

Coordinate human-AI pair programming sessions with Driver, Navigator, and Switch modes.

Official
Intermediate