Skip to main content

code-review

Review current branch changes against REVIEW.md guidelines

来源信息

仓库
TheBushidoCollective/han
最近来源活动
2026年8月12日 09:21
检测到的 SKILL.md 语言
英语
星标
198
Forks
24

安装方式

默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。

检查来源文件

决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。

正在显示 SKILL.md

SKILL.md
来源说明 · 只读预览
name
code-review
description
Review current branch changes against REVIEW.md guidelines
# Code Review ## Name session-code-review:code-review - Review session changes against REVIEW.md and CLAUDE.md guidelines ## Synopsis ``` /code-review [--branch <base>] ``` ## Description Performs a thorough code review of all changes on the current branch compared to the base branch (defaults to `main` or `master`). Reviews against `REVIEW.md` and `CLAUDE.md` guidelines in the repository. This is a local-only review — no PR or GitHub integration required. It examines your uncommitted changes, staged changes, and all commits on the current branch. ## Implementation Follow these steps exactly: ### 1. Determine the base branch If `--branch` was provided, use that. Otherwise, detect the default branch: ```bash git remote show origin | grep 'HEAD branch' | sed 's/.*: //' ``` ### 2. Gather review guidelines Read the following files if they exist: - `REVIEW.md` at the repo root - `CLAUDE.md` at the repo root - Any `CLAUDE.md` files in directories containing changed files These define the review criteria. If no `REVIEW.md` or `CLAUDE.md` exists, use general best practices (security, correctness, logic errors). ### 3. Get the full diff ```bash # Uncommitted changes git diff # Staged changes git diff --cached # All commits on this branch vs base git diff <base-branch>...HEAD ``` ### 4. Review the changes For each changed file, check against the guidelines from step 2. Focus on: - **REVIEW.md violations**: Explicit rules the team has defined - **CLAUDE.md violations**: Project conventions and requirements - **Bugs**: Logic errors, off-by-one errors, null pointer issues - **Security**: Injection, auth bypass, exposed secrets, OWASP top 10 - **Correctness**: Code that will produce wrong results Do NOT flag: - Style preferences (unless in REVIEW.md) - Issues a linter catches - Subjective improvements - Pre-existing issues not in the diff ### 5. Report findings For each issue found, report: - **File and line number** (link format: `file_path:line_number`) - **Severity**: critical / warning - **Description**: What's wrong and why - **Suggestion**: How to fix it If no issues are found, report: "No issues found. Changes comply with REVIEW.md and CLAUDE.md guidelines." ### 6. Offer to fix If issues were found, ask the user if they'd like you to fix them. ## Example Interaction ``` User: /code-review Claude: Reviewing changes on branch `feature/auth-flow` against `main`... Found REVIEW.md with 3 rules: 1. All API routes must validate auth tokens 2. Database queries must use parameterized statements 3. Error responses must not leak internal details Reviewing 4 changed files... ## Code Review Results ### 1. Missing auth validation (critical) `src/routes/users.ts:45` — New GET `/users/:id` endpoint does not validate the auth token. REVIEW.md rule: "All API routes must validate auth tokens." **Fix**: Add `validateToken(req)` middleware before the handler. ### 2. SQL injection risk (critical) `src/db/queries.ts:23` — String interpolation in SQL query: `` `SELECT * FROM users WHERE id = ${id}` ``. REVIEW.md rule: "Database queries must use parameterized statements." **Fix**: Use `db.query('SELECT * FROM users WHERE id = ?', [id])`. --- Found 2 issues (2 critical). Would you like me to fix them? ```
在 GitHub 查看