By LouisLin0723
Force the AI to rigorously self-review its own work before shipping — catches the bugs, holes and hallucinations it missed the first time. Bidirectional review + reverse rebuttal + domain checklists. Includes 1/2/3 shortcut workflow.
逼 AI 把自己刚做的事往死里复盘一遍,揪出它第一遍没看见的 bug、漏洞、幻觉。 Force the AI to review its own work — catch the bugs, holes & hallucinations it missed.
每个用 AI 写代码的人都中过招:AI 信誓旦旦说"搞定了",结果一跑全是 bug。这不是它态度差,是模型的先天限制 —— 它有幻觉率 (hallucination),而且一次看不全。
把 AI 想象成一个只有两只眼睛的人。 它盯着面前这 1,000 个文件,就只能看到这 1,000 个。 但真正的 bug,往往不在它眼皮底下这 1,000 个里 —— 而是藏在它没加载进来的另外 99,000 个文件里。 整个项目有 10 万个文件,没有任何 AI 能一口气全看完。 它看了 1% 就敢喊"搞定了",剩下 99% 的雷,它根本没瞧见。
Picture the AI as a person with only two eyes. Staring at the 1,000 files in front of it, it sees only those 1,000. The real bugs are hiding in the other 99,000 files it never loaded. It says "done" after seeing 1% — and 99% of the landmines were never in view.
double-check = 再花一点 token,逼 AI 回头把自己刚做的事仔细复盘一遍。靠这套亲测有效的提示词,它能再揪出一大批第一遍漏掉的洞。
flowchart TD
A["AI 说: 搞定了 ✅<br/>AI: it's done"] --> B{double-check}
B --> C["① 双向推演<br/>0→100 顺推 + 100→0 反推"]
B --> D["② 逆向反驳<br/>逼 AI 找自己的漏洞"]
B --> E["③ 分领域清单<br/>功能 13 / 代码 21 / 算法 18 问"]
C --> F["🐛 揪出第一遍漏掉的<br/>bug · 漏洞 · 幻觉"]
D --> F
E --> F
style B fill:#DC2626,color:#fff
style F fill:#16A34A,color:#fff
复盘消不掉幻觉(没人能),但能把 AI 的严谨度实打实再顶高一档。
你让 Claude 写个「取数组中位数」的函数,它给你:
function median(arr) {
arr.sort((a, b) => a - b);
return arr[Math.floor(arr.length / 2)];
}
Claude:"搞定了 ✅"
你打 2。double-check 用代码清单逼它复盘自己,当场揪出 3 个它第一遍没看见的洞:
arr.sort() 原地改了调用者传进来的数组,留下副作用median([]) 返回 undefined,根本没处理第一遍那句"搞定了"漏了 3 个 bug。这就是 double-check 的价值 —— 把"看了 1% 就喊完成"逼成"再查一遍"。
Ask Claude for a
median()function and it confidently ships one with three bugs: even-length midpoint, in-placesort()side-effect, and an empty-array crash. Hit2— double-check's checklist surfaces all three. That's the point.
复盘是核心,触发它只要一个动作:
| 你做什么 | 会发生 |
|---|---|
打 2,或说"复盘 / double-check / 审一遍 / 有没有问题" | AI 立刻进入严格复盘:双向推演 + 逆向反驳 + 分领域清单 |
| 涉钱 / 不可逆 / 重大决策 | 自动触发复盘,不用你喊 |
附带 1 / 2 / 3 快捷键工作流:
1=干到底、2=复盘(核心)、3=上线收尾。快捷键是辅助,复盘才是主角。完整方法论见SKILL.md。
A. Plugin 一键装(推荐):
/plugin marketplace add LouisLin0723/double-check
/plugin install double-check@double-check
B. 手动 clone:
git clone https://github.com/LouisLin0723/double-check.git
cp -r double-check/skills/double-check ~/.claude/skills/
重启 Claude Code 生效。
double-check 强制它反向找茬,戳穿"虚假完成"。复盘清单 + 快捷键是作者的工作流约定。Fork 它,改 SKILL.md 里的触发词、检查清单、部署命令,换成你自己的节奏。
不绑定任何框架 —— 一套可迁移的复盘骨架。 Framework-agnostic. A portable self-review skeleton.
MIT © Louis Lin
每天被 AI 的"搞定了"坑过? 点个 ⭐ Star,下次让它先 double-check ❤️ Tired of your AI's fake "done"? Drop a ⭐ — make it double-check next time.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub louislin0723/double-check --plugin double-checkA growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
UI/UX design intelligence. 67 styles, 161 palettes, 57 font pairings, 25 charts, 15 stacks (React, Next.js, Vue, Svelte, Astro, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui, Nuxt, Jetpack Compose). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.