Skip to main content

code-review-personal

Perform a comprehensive branch code review using `git diff main...HEAD`, feature-plan context files, and project conventions. Use when a user asks for a personal/manual PR review, quality gate, or merge-readiness assessment with severity-ranked findings and a final verdict.

ソース情報

リポジトリ
sunpar/agentic_coding_tools
ソースの最終更新活動
2026年2月13日 08:39
検出された SKILL.md の言語
英語
スター
1
フォーク
0

インストール方法

デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。

ソースファイルを確認

インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。

ファイルエクスプローラー
2 ファイル

SKILL.md を表示中

SKILL.md
ソースの指示 · 読み取り専用プレビュー
name
code-review-personal
description
Perform a comprehensive branch code review using `git diff main...HEAD`, feature-plan context files, and project conventions. Use when a user asks for a personal/manual PR review, quality gate, or merge-readiness assessment with severity-ranked findings and a final verdict.
# Code Review Personal ## Workflow 1. Run `git diff main...HEAD` to inspect all branch changes. 2. Load context from `FEATURE_PLAN.md` or `feature_summary.md` if either exists. 3. Load context from `ORIGINAL_PLAN.md` if it exists. 4. Review all changes against the criteria below. ## Review Criteria ### Functionality - Verify code matches feature intent from the plan files. - Check edge-case handling. - Check error handling appropriateness. - Flag obvious bugs or behavioral regressions. ### Code Quality - Check readability, consistency, and organization. - Prefer single-purpose functions. - Check naming clarity. - Enforce project conventions from `CLAUDE.md`. ### Security and Privacy - Flag any sensitive data logging (raw transactions, account numbers, CSV rows). - Check input validation where needed. - Flag hardcoded secrets or credential-like values. ### Performance - Flag inefficient algorithms or query patterns. - Flag potential N+1 access patterns. - Flag unnecessary recomputation or heavy work. ### Tests - Check whether tests cover new behavior and risk areas. - Prefer behavior-focused tests over brittle implementation-coupled tests. - If additional tests are needed, use the `/write-tests` skill. ## Output Format Provide all sections in this order: 1. Executive Summary (1-2 paragraphs) 2. Findings by Severity (Critical > Major > Minor > Cosmetic) 3. Test Coverage Assessment 4. Final Verdict: `Ready to merge` | `Needs minor cleanup` | `Needs major revisions` ## Ground Rules - Prioritize findings and risks over praise. - Keep findings specific and actionable. - Reference concrete file paths and lines when possible. - If no findings exist, state that explicitly and note residual risk/testing gaps.
GitHubで見る