skills.rest
HACK-WU
HACK-WUCOMMUNITY
Explore all skills by this author →
GitHub

code-review

Review git diffs and commits across seven dimensions with semantic consistency analysis.

2|Updated Jun 8, 2026
Search Tags:#code review#pull request#security audit#multi-language#semantic analysis#git diff#merge review
One-click install
npx skills add https://github.com/HACK-WU/skills --skill code-review-hack-wu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: code-review
Source: https://github.com/HACK-WU/skills/tree/main/skills/code-review
Command: npx skills add https://github.com/HACK-WU/skills --skill code-review-hack-wu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Code can be technically correct yet implement the wrong thing. This Skill catches semantic drift between stated intent and actual implementation, alongside security, bug, style, architecture, performance, and test-coverage issues in code changes. ## Core Features & Use Cases - Seven-dimension review: Semantic consistency, security, bug risk, code style, architecture, performance, and test coverage, with review depth adapting to change size (quick/standard/deep modes). - Intent inference and confirmation: Infers the functional intent of changed code from naming, docstrings, callers, and linked requirements, then verifies implementation against that intent. - Structured reports and follow-up: Produces a scored review report with severity-ranked findings, auto-triggers adversarial re-checks on clean passes, and runs temporary tests to validate uncertain data flows. - Use Case: Paste a commit hash or run it on a pull request to receive a full multi-dimensional review report with line-level findings and merge recommendations. ## Quick Start Ask the AI to review a specific commit, branch diff, or pull request, for example by saying: review the changes in commit abc1234.

Frequently Asked Questions about code-review

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

FAQPage Schema
How do I review a git commit or pull request with AI?▼

Provide a commit hash, git diff output, branch comparison, or PR number, and the review fetches the diff using git or the GitHub CLI. It then reads the full changed files, infers intent, and produces a structured multi-dimension report.

What does a semantic consistency code review check?▼

It verifies that the implementation matches the stated intent from commit messages, docstrings, naming, and linked requirements. It also checks whether local changes break the overall semantics of the enclosing method, class, module, or data flow.

Which programming languages does this code review support?▼

It supports Python, JavaScript, TypeScript, Java, Go, and Rust. The review adapts to each language's idioms, common pitfalls, and declared runtime versions detected from project configuration files.

Can I review code without a git repository context?▼

Yes, you can paste code changes directly. The review skips commit-message and granularity checks, marks the report as having no commit context, and may ask for the containing file or function for better analysis.

How does review depth change with the size of the change?▼

Changes of five lines or fewer use a quick mode covering semantics, security, and bugs only. Larger changes trigger standard or deep modes covering all seven dimensions, with parallel sub-reviews and architecture impact analysis.

What happens when a review finds no blocking issues?▼

For standard-mode reviews with a clean pass, an adversarial challenger re-examination runs automatically to guard against reviewer blind spots. If it finds blocking concerns, the merge recommendation is updated accordingly.

Related Skills

Explore More Software Engineering Skills→
EveryIncEveryInc
24.7k

ce-commit

Create git commits with clear, outcome-focused messages from staged or unstaged changes.

Official
Intermediate
seqisseqis

agent-peer-review-checker

Verifies code review compliance by tracking comments and analyzing diffs for blockers and force-pushes.

Community
Advanced
SilentFlowerSilentFlower

trellis-run-full-chain

Executes cross-layer end-to-end verification scenarios across UI, API, and database layers.

Community
Advanced
managedcodemanagedcode
52

mcaf-formatting

Apply repository-defined formatter and linting commands to changed code.

Official
Basic