菜单
Use when the user asks for a code review, "review this change", "review my PR", "review the diff", or wants quality/spec/security/perf feedback on recent changes. Triggers a multi-level review with a thinking-tier reviewer agent. Standalone — does not auto-chain.
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。
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。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/ccplugins/awesome-claude-code-plugins --skill audit命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Find screenshots shared/accumulated during this Claude Code session and, with the user's approval, move them to the Trash (no permanent delete — recoverable). Use before ending a session, when asked to tidy the Desktop/screenshots folder, or when the screenshot-janitor Stop hook suggests cleanup. Triggers include "clean up screenshots", "delete screenshots", "tidy my desktop", "remove the screenshots".
Saves and searches past Claude Code sessions so context, decisions, and code persist across conversations. Use when user says 'remember this', 'save this session', 'recall', 'search past sessions', 'what did we discuss about', or 'find previous work on'. Do NOT use for general file storage, note-taking, or bookmark management.
Use this skill when the user asks for a tarot reading, I Ching (Yijing) hexagram cast, or Chinese Xiao Liu Ren (小六壬) divination. Calls audited local Python tools that produce deterministic JSON output; the model never invents the result, it only interprets the JSON.
基于 SOC 职业分类
正在显示 SKILL.md
| name | audit |
| description | Use when the user asks for a code review, "review this change", "review my PR", "review the diff", or wants quality/spec/security/perf feedback on recent changes. Triggers a multi-level review with a thinking-tier reviewer agent. Standalone — does not auto-chain. |
Standalone multi-level code review. Dispatcher — Opus 4.7 (thinking-tier). Workers — Sonnet 4.6.
git diff HEAD + git diff --staged--level 1 through --level 5 (default — L2)Adapted from review-levels.md:
| L | Name | Checks |
|---|---|---|
| 1 | Quick | Syntax, obvious bugs, formatting |
| 2 | Standard | L1 + spec compliance, naming, edge cases |
| 3 | Thorough | L2 + cross-file consistency, integration risks, security |
| 4 | Deep | L3 + architecture, scalability, accessibility |
| 5 | Exhaustive | L4 + adversarial probing, perf profiling, alternatives |
Security scan (hardcoded secrets, injection, path traversal, XSS, missing validation) is mandatory at L3+. See security.md.
git diff HEAD + git diff --staged.Searcher — gathering context for review (Sonnet 4.6) to map referenced files and load relevant project context.**Reviewer** — reviewing <scope> at level L<n> — Opus 4.7 (thinking-tier, non-negotiable)..hyperflow/memory/learnings.md per memory-system.md.If any security issue found → emit SECURITY_VIOLATION: halt marker immediately.
── Review Result ──────────────────────
Scope: <files / range / commit>
Level: L<n>
Verdict: PASS | NEEDS_FIX | SECURITY_VIOLATION
[Critical]
- file:line — issue + required fix
[Important]
- file:line — issue + recommended fix
[Suggestions]
- file:line — optional improvement
[Praise]
- file:line — what's done well
───────────────────────────────────────
Agents: 1 searcher (sonnet) · 1 reviewer (opus)
/hyperflow:deploy if the user is ready to release. Do not auto-ship./hyperflow:trace (for root-cause bugs) or manual edits. Do not auto-fix.Full rules in DOCTRINE.md. Output style in output-style.md.