REA
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
Architecture diagrams and checkable docs from your codebase — ERD, C4, API and lifecycle — generated from Prisma, OpenAPI or GraphQL. One HTML file, no server.
如果你是作者,可把这段代码贴到 README。徽章会链接回本页。
[](https://mcp.popnx.com/en/skills/raja0sama-vibex)在同一工作中把 AI 连到服务的 MCP
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
Codebase intelligence for TypeScript and JavaScript, exposed as structured MCP analysis tools.
MCP server for terminal commands, file operations, and process management
同类工作,排序与 Skill 列表相同
| 本项 vibex | rea | andrej-karpathy-skills | graphify | MoneyPrinterTurbo | |
|---|---|---|---|---|---|
| 作者 | Raja0sama | morluto | multica-ai | Graphify-Labs | harry0703 |
| Skill 数 | 1 | 1 | 1 | 1 | 1 |
| GitHub 星标 | 392 | 465,460 | 233,910 | 149,364 | 144,149 |
| 最近代码更新 | 2026年11月11日 | 2026年11月17日 | 2026年4月20日 | 2026年11月17日 | 2026年11月17日 |
| 许可证 | MIT | MIT | – | Apache-2.0 | MIT |
Raja0sama 提供的 1 个 Skill(指令集),AI 读取后照做 · 不是 MCP server
MCP = 手
把 AI 连到外部服务或数据,例如读取 Notion 或开发票。需要有程序或服务器在运行。
Skill = 手册
教 AI 怎么做某项工作的文本文件,不连接任何服务。
有些会重叠:少数 Skill 带有调用工具的脚本,少数 MCP 只提供知识。请从你想完成的工作出发。
Claude CodeCodexCursorGemini CLIGitHub CopilotOpenCode
依据开发者 README 和 repo 中的文件夹
GitHub account Raja0sama (person), 9 years old
★ 392 stars, few users so far · code last changed 2026年11月11日
60 scripts that can run on a computer: bin/vibex.mjs, assets/viewer.js, examples/shop-repo/bff/src/auth/identity.interceptor.ts, examples/shop-repo/bff/src/auth/session.service.ts and 56 more
Scripts that go online: examples/shop-repo/bff/src/orders/orders.client.ts
No hooks or writes to machine settings found
Checked automatically from the repo files at commit 0f60147 (read only, nothing was run). Not a safety guarantee: read it and decide for yourself
第一步:你用哪个 AI?
选择你用的应用,本页只显示该应用的安装方法
Uploading to claude.ai would fail: the description is over 200 characters (976). Use "Let the AI install it" instead
复制这句话,粘贴到 Claude Code 的对话中
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
重新打开 Claude Code,输入 /vibex,或直接提出相关任务,Claude 会自动选用
Checked against official docs 2026年11月12日 · Source: code.claude.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
已经装好了吗?先检查 · 打开终端,输入这条命令后按 Enter
node -vMac:按 ⌘ + Space,输入 Terminal 后按 Enter
Windows:打开“开始”菜单,输入 Windows PowerShell,然后选“打开”
还没安装: 打开官方下载页,选择 LTS 版本和点选式安装程序:Mac 用 .pkg · Windows 用 .msi。双击后一路下一步即可,不需要 curl 命令
装好后关掉终端再打开,重新运行同一条命令,应显示版本号: node -v
安装一次,所有需要 Node.js 的工具都能用
正在检查命令 · 通过后才会显示复制按钮
摘自开发者 README,会下载并运行开发者的代码
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
ChatGPT 桌面应用内置 Codex,Codex 从 ~/.agents/skills 文件夹读取 Skill
在 ChatGPT 桌面应用中打开 Codex,把这句话粘贴到对话中
The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
把 vibex 文件夹放到 ~/.agents/skills/vibex
~ 是你的主文件夹($HOME)
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
不能使用你自己安装的 Skill,只能使用目录中插件自带的 Skill
想在 ChatGPT 中使用这个 Skill,请按上面的方法使用 ChatGPT 桌面应用
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
复制这句话,粘贴到 Codex 的对话中 · 适用于 ChatGPT 桌面应用中的 Codex、Codex CLI 和 IDE 扩展
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Codex 会自动发现新安装的 Skill;如果没有出现,请重启 Codex · 使用方法:在 ChatGPT 中输入 @ 并选择 Skill · 在 Codex CLI 或 IDE 扩展中运行 /skills,或输入 $ 加 Skill 名称
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
复制这句话,粘贴到 Cursor 的对话中
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
在 Cursor 的 Agent 聊天中输入 / 并选择 vibex
Checked against official docs 2026年11月12日 · Source: cursor.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
适用于其他能读取 Skill 的应用。这句话列出了 Claude Code、Codex 和 Cursor 的路径,AI 会按你的应用选择 · Copy this sentence into the app’s chat
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Checked against official docs 2026年11月12日 · Source: code.claude.com, learn.chatgpt.com, cursor.com
其他支持 Agent Skills 的应用(如 Gemini CLI、GitHub Copilot)使用不同文件夹,请查看该应用的指南,再把含 SKILL.md 的文件夹复制过去。
已于 2026年11月11日 对照 Claude Code、Codex 和 Cursor 官方指南核实 skill 文件夹
以下文字来自开发者的 repo,不是 PopNx MCP 的指令
显示前 8,000 个字符 · 在 GitHub 打开完整文件
---
name: vibex
description: 'Diagrams and checkable docs from a codebase. Draws ERDs, C4 context/container/component views, API endpoint catalogues, lifecycle/state machines, sequence diagrams, and service-to-service call maps as a validated JSON spec plus a standalone HTML viewer, with Mermaid export. Imports OpenAPI, GraphQL SDL and Prisma; otherwise reads the code (NestJS, TypeORM, Express) and writes the spec. Use for: database diagram, ERD, data model, table relationships, C4 or architecture diagram, system context, API map, endpoint list, "show me the endpoints", status flow, state machine, allowed transitions, "what happens after X is approved", sequence diagram, request flow, how services call each other, microservice map, Mermaid diagram. Also writes documentation as a fact graph: each statement carries its source and a computed confidence, and a drift check fails CI when the code moves. Use to document a system, write architecture or onboarding docs, or explain how a system works.'
license: MIT
metadata:
version: "0.8.1"
cli: node bin/vibex.mjs
repository: https://github.com/Raja0sama/vibex
---
# vibeX
*You vibed it into existence. vibeX shows you what you actually built.*
Output is always two files: `<name>.<type>.json` (the spec, the source of truth) and `<name>.<type>.html` (self-contained viewer: pan/zoom, search, click-for-details, dark/light, SVG/PNG export). Not fancy. Correct and readable.
Skill root: the directory containing this file. Every command below is `node <skill-root>/bin/vibex.mjs …`.
## Steps
1. **Pick the type** from the ask. One diagram per type per question; do not mix.
| Ask | Type | Spec file |
|---|---|---|
| tables, models, schema, relations, FKs, data model | `erd` | `schemas/erd.schema.json` |
| context, containers, components, who talks to what, system boundary | `c4` | `schemas/c4.schema.json` |
| endpoints, routes, API surface, GraphQL operations, events/topics | `endpoints` | `schemas/endpoints.schema.json` |
| statuses, state machine, lifecycle, what happens after X, allowed transitions | `lifecycle` | `schemas/lifecycle.schema.json` |
| sequence diagram, request flow, what happens when the user clicks X, who calls whom and in what order | `sequence` | `schemas/sequence.schema.json` |
2. **Find a machine-readable source first.** Search the repo before reading code:
- `endpoints`: `openapi.*`, `swagger.*`, `*.graphql`, `schema.gql`, or a generated schema endpoint dump. Run `import openapi` or `import graphql`.
- `erd`: `schema.prisma` → `import prisma`. A GraphQL SDL can also seed an ERD with `import graphql <file> out.json --erd`.
- `c4`: never importable. Author it (step 3).
```bash
node bin/vibex.mjs import openapi path/to/openapi.yaml out/api.endpoints.json
node bin/vibex.mjs import graphql path/to/schema.graphql out/api.endpoints.json
node bin/vibex.mjs import prisma prisma/schema.prisma out/db.erd.json
```
After an import, open the JSON and edit it like a human would: rename groups, drop noise endpoints (health, metrics), add `entities` links on endpoints, add `sources`. The import is a starting point, not the deliverable.
3. **No source file? Dig in the code.** Read the schema file for your type once, then read one example in `examples/`. Author the JSON fresh. Where to look:
- **NestJS REST**: `@Controller('prefix')` + `@Get/@Post/@Put/@Patch/@Delete('path')` → one endpoint each, path = prefix + path with `:id` rewritten as `{id}`. `@UseGuards(...)`/`@Roles(...)` → `auth`. `@Body()` DTO class → `request`. Return type / `@ApiOkResponse` → `response`. One group per controller. Put `sources: [{path, line}]` on every endpoint pointing at the handler method. Prepend `app.setGlobalPrefix()` and `@Version()`/`VersioningType.URI` segments to every path. `@Param()` → `params[].in: "path"`, `@Query()` → `"query"`, `@Headers()` → `"header"`; a DTO class in `@Query()` becomes one param per property.
- **NestJS GraphQL**: `@Query()`, `@Mutation()`, `@Subscription()` in `*.resolver.ts` → method QUERY/MUTATION/SUBSCRIPTION, path = `name(arg: Type)`. `@ObjectType`/`@InputType` classes → `types`.
- **Express/Fastify**: `router.get('/x', …)`, `app.post(...)`, `fastify.route({method, url})`.
- **TypeORM**: `@Entity('table')` classes; `@PrimaryGeneratedColumn`/`@PrimaryColumn` → `pk`; `@Column({nullable, unique, type})`; `@ManyToOne` + `@JoinColumn` → FK on this side, relationship `from` this entity `to` target with `from_cardinality: many`; `@OneToOne` → `one`/`zero-or-one`; `@ManyToMany` + `@JoinTable` → a join entity or a many↔many relationship.
- **Prisma**: prefer the importer. The importer already fills `sources` (path + model line); add `description` and `groups` by hand. It names entities by the lowercased model name (`OrderItem` → `orderitem`; `@@map` only changes the label).
- **SQL migrations**: `CREATE TABLE`, `REFERENCES`, `PRIMARY KEY`, `UNIQUE`.
- **C4**: `docker-compose.yml`, `k8s/`, `serverless.yml`, `infra/` for containers and datastores; `package.json`/`*.module.ts` for tech; HTTP clients, queue clients, SDK imports (`stripe`, `@aws-sdk/client-sqs`, `nodemailer`) for external systems and relationships. People come from auth roles.
Keep IDs stable and boring: `user`, `order_item`, `bff`, `list-orders`. Reuse the same ID for the same thing across diagrams so ERD entity IDs can go into `endpoints[].entities`. With a Prisma-imported ERD, use exactly the importer's ids (`orderitem`, not `order_item`); run `dashboard` and check the "which endpoints touch which tables" table is non-empty.
4. **Validate, then render.**
```bash
node bin/vibex.mjs validate out/db.erd.json
node bin/vibex.mjs render out/db.erd.json out/db.erd.html
```
Exit 1 means errors: fix the named field and rerun. Warnings never block; fix the ones about layout (`row`/`col` hints) when they appear, ignore missing-summary warnings unless the user wants prose. Add `--open` to the render to open the browser.
5. **Several diagrams for one system? Build the dashboard too.** Keep all specs in one folder (e.g. `docs/arch/`), then:
```bash
node bin/vibex.mjs dashboard docs/arch/index.html docs/arch --title "My system"
```
The overview page lists every diagram, shows which endpoints touch which tables (from `endpoints[].entities`), and resolves C4 `link` values that name another spec file in the folder (`link: "bff.c4.html"` → the `bff.c4.json` panel; both specs must be in the same `dashboard` invocation, matching is by file name). Re-run it after every spec change; it is cheap.
**Several repos side by side?** Run from the folder above them. Keep each repo's specs inside that repo (e.g. `repo-a/docs/arch/`) so its CI can check drift; a shared folder of cross-repo specs works too, with anchor paths prefixed by the repo folder (`repo-a/src/x.ts`). Then:
```bash
node bin/vibex.mjs workspace ~/work --title "My platform" # → ~/work/vibex-site/
node bin/vibex.mjs workspace ~/work --host # + Dockerfile, nginx, login
```
One dashboard per top-level folder that holds specs, plus a landing page listing them. Re-run after any spec change; folders whose specs are gone are removed from the site.
6. **Report**: the two paths, counts (entities/elements/endpoints and relationships), what was imported vs inferred, and anything you left out on purpose. If code reading left ambiguity (cardinality, auth), say so in one line and put it in a `cards` note with `tone: "warning"`.
## Documentation (`docs`)
A `docs` spec is not a diagram. Its unit is a **claim**: one checkable sentence with a source. The build turns claims plus the diagram specs into `docs.json` — the fact graph humans read in the dashboard and agents read directly.
```bash
node bin/vibex.mjs docs docs/arch/system.docs.json docs/arch --repo . --reanchor # pin anchors
node bin/vibex.mjs docs docs/arch/system.docs.json提问、分享技巧或报告无法使用。按“有帮助”数量排序。