[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-80208":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":14,"stars30d":15,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":17,"fork":17,"defaultBranch":18,"hasWiki":19,"hasPages":17,"topics":20,"createdAt":9,"pushedAt":9,"updatedAt":21,"readmeContent":22,"aiSummary":23,"trendingCount":14,"starSnapshotCount":14,"syncStatus":24,"lastSyncTime":25,"discoverSource":26},80208,"TigerAI-n8n-Skill-Pack","MorrisLu-Taipei\u002FTigerAI-n8n-Skill-Pack","MorrisLu-Taipei","Build enterprise-grade n8n workflows using natural language sticky notes: An AI engine that transforms simple human intent into complete, documented three-layer automation systems.",null,"JavaScript",72,18,1,0,3,3.84,false,"main",true,[],"2026-06-12 02:03:59","# TigerAI n8n Skill Pack — User Manual\n\n> 🌐 **English** | [繁體中文](README.zh.md)\n\n> Describe what you want in plain language (like talking to a coworker), and AI generates a complete n8n workflow for you.\n> No coding required.\n\n![TigerAI n8n Skill Pack — full pipeline diagram](docs\u002Fimages\u002Ftigerai-flow-en.png)\n\n> 📊 **The whole pack in one picture**: User writes yellow sticky notes (Layer 1 intent) → TigerAI Skill Pack brain (Cookbook + 2,061 reference workflows + DSL v1.2 + 12 skills + 4 enterprise patterns) → Three-layer workflow JSON (real n8n workflow). Three features, zero learning curve.\n> *by n8n Taipei Ambassador Morris Lu*\n\n---\n\n## 🤖 This is an Agentic Engineering Example\n\n> **This entire project was authored using AI Agentic IDEs (Antigravity \u002F Claude Code) — from spec to n8n workflows, every artifact was produced through human-AI agent collaboration.**\n\nThis Skill Pack is itself a working demo of **Agentic Engineering**:\n\n| Dimension | Traditional way | This project (Agentic) |\n|---|---|---|\n| **Spec writing** | Engineer types every word | Chat with AI → AI produces SDD (Spec-Driven Design) |\n| **n8n workflow dev** | Drag nodes on canvas | Write a yellow sticky note → AI emits runnable JSON |\n| **Skill \u002F plugin authoring** | Read docs, copy templates | Claude Code Skills + Antigravity `.agent\u002Fworkflows\u002F` orchestration |\n| **Acceptance testing** | Run cases by hand, write report | AI runs 8 scenarios → auto-emits [`tests\u002FREPORT-3.en.md`](tests\u002FREPORT-3.en.md) |\n| **Docs \u002F README \u002F CHANGELOG** | Backfilled after coding | Generated alongside code |\n| **Third-party license compliance** | Manual review | AI detects leaked secrets, scrubs them, generates `THIRD_PARTY_NOTICES.md` |\n\n### Agentic footprints in this repo\n\n- **`skills\u002F`** — 13 Claude Code \u002F Antigravity skills; each `SKILL.md` is co-authored by humans and AI\n- **`.agent\u002Fworkflows\u002F`** — Antigravity-native agentic workflows (e.g. `\u002Finstall-n8n-pack` one-shot installer)\n- **`cookbook\u002F`** — 8 natural-language → workflow examples showing how to \"talk to\" the AI\n- **`spec\u002Fsticky-note-three-layer.md`** — Three-layer structure spec that forces reviewable AI output\n- **`research\u002Fpatterns.md`** — 7 canonical skeletons + anti-patterns mined by AI from 2,061 real workflows\n- **`reference-workflows\u002F`** — AI training corpus ([Zie619\u002Fn8n-workflows](https:\u002F\u002Fgithub.com\u002FZie619\u002Fn8n-workflows), MIT, secrets scrubbed)\n\n### Who should study this project\n\n- Developers \u002F PMs learning **how to use an AI agent as an engineering teammate**\n- Teams evaluating **whether Antigravity \u002F Claude Code can replace hand-written skills \u002F workflows**\n- Anyone curious **what real human-AI co-authored engineering output looks like**\n\n> 💡 In other words: this isn't just \"a Skill Pack for n8n\" — it's also an open **case study of how AI agents build a real product**.\n\n### 👥 You (the user) can build n8n workflows the same way\n\n**Once you install this Skill Pack, you can author your own n8n workflows with the same agentic approach** — no node syntax to learn, no code to write:\n\n| Tool | What you do | What the AI does |\n|---|---|---|\n| **Antigravity** | Open your n8n project in Antigravity, run `\u002Finstall-n8n-pack`, then describe what you want in plain language | `.agent\u002Fworkflows\u002F` auto-reads your intent → emits workflow JSON → deploys via n8n API |\n| **Claude Code (CLI \u002F VS Code)** | Run `bash install.sh` (or `install.ps1`) in your working dir, then say \"I want a workflow that…\" | 13 skills auto-load → three-layer workflow produced → ready to import into n8n |\n| **Any AI assistant (ChatGPT \u002F Gemini)** | Paste an example from [`cookbook\u002F`](cookbook\u002F00-INDEX.en.md) as a few-shot prompt | Imitates the three-layer structure and emits a compliant workflow JSON |\n\n**Typical interaction** (30-second mental model):\n\n```text\nYou ──> AI: \"Every weekday 9am, pull Shopify orders, build a daily\n             report, email it to the boss; on failure post to Slack #ops\"\n\nAI ──> You: ✅ workflow.json generated (Schedule → Shopify → Code → Email + Error → Slack)\n             ✅ Yellow sticky: your original requirement, preserved\n             ✅ Blue sticky: which credentials, constraints, test method\n             ✅ Deployed to your n8n via API, webhook URL: https:\u002F\u002F...\n```\n\n> 🎯 **The core idea**: Users don't need to understand n8n internals — they just need to \"talk like a human\" to the AI. The Skill Pack ensures the AI's output is spec-compliant, reviewable, and maintainable.\n\nSee [`02-USAGE-MODES.en.md`](02-USAGE-MODES.en.md) for the three usage modes and [`03-FIRST-WORKFLOW.en.md`](03-FIRST-WORKFLOW.en.md) for a 15-minute hands-on walkthrough.\n\n---\n\n## 📖 Reading order (strongly recommended)\n\n| # | File | Audience \u002F Time |\n|---|---|---|\n| 0️⃣ | **This README.md** | Overview, start here (5 min) |\n| 1️⃣ | [`01-INSTALL.en.md`](01-INSTALL.en.md) | First-time setup (10 min) |\n| 2️⃣ | [`02-USAGE-MODES.en.md`](02-USAGE-MODES.en.md) | Pick your usage style (5 min) |\n| 3️⃣ | [`03-FIRST-WORKFLOW.en.md`](03-FIRST-WORKFLOW.en.md) | Hands-on: build your first workflow (15 min) |\n| 4️⃣ | [`04-FAQ.en.md`](04-FAQ.en.md) | Reference when stuck |\n\n---\n\n## ⚡ Understand it in 90 seconds\n\n### What it does\n\nYou drop a **yellow sticky note** on the n8n canvas and write (in any language):\n\n```text\nEvery day at 9 AM, fetch sales data and email the daily report to my boss.\nOn failure, notify Slack #ops.\n```\n\nYou ask AI to build it. The canvas now shows a complete workflow:\n\n```\n┌─ Yellow sticky: your requirement (preserved as-is)\n├─ Middle: AI-generated nodes (Schedule → HTTP → Code → Email)\n└─ Blue sticky: AI's notes (credentials needed, assumptions, limitations, how to test)\n```\n\nNo code. No syntax to learn. No need to memorize n8n node names.\n\n### Three usage modes (details in [02-USAGE-MODES.en.md](02-USAGE-MODES.en.md))\n\n| Mode | When | Trigger phrase |\n|---|---|---|\n| 🪄 Cookbook copy | You know what you want, fast | Copy from [cookbook](cookbook\u002F00-INDEX.en.md) |\n| 💬 Q&A mode | You have no idea how to describe it | \"enable Q&A mode\" \u002F \"問答模式\" |\n| 🔍 Example finder | Want to see prior art first | \"find examples for X\" \u002F \"範例查詢\" |\n\n---\n\n## 📂 Pack contents\n\n```text\nTigerAI-n8n-Skill-Pack\u002F\n├── README.md \u002F README.zh.md ← You are here\n├── 01-INSTALL.md\u002F.en.md       ← Install\n├── 02-USAGE-MODES.md\u002F.en.md   ← Three usage modes\n├── 03-FIRST-WORKFLOW.md\u002F.en.md ← Hands-on tutorial\n├── 04-FAQ.md\u002F.en.md           ← Common questions\n│\n├── cookbook\u002F                  ← 8 copy-paste recipes (each has plain-language + DSL fold)\n│   └── 00-INDEX.md\u002F.en.md\n│\n├── skills\u002F                    ← 13 Skills loaded by AI assistant\n│   ├── _vendor\u002F                  7 official n8n-skills (MIT)\n│   └── tigerai\u002F                  6 TigerAI custom (incl. AG Auto-Install)\n│\n├── spec\u002F                      ← Technical specs (for engineers)\n├── examples\u002Ftigerai-flagship\u002F ← 3 enterprise-grade examples (with SDD)\n├── reference-workflows\u002F       ← 2,061 public workflows (AI corpus)\n├── research\u002F                  ← Research artifacts\n├── tests\u002F                     ← Three rounds of acceptance reports\n│\n├── CHANGELOG.md \u002F VERSION\n├── install.sh \u002F install.ps1   ← Install scripts (Supports Claude Code & Antigravity)\n├── .agent\u002Fworkflows\u002F          ← Antigravity-exclusive workflows (e.g., \u002Finstall-n8n-pack)\n└── plugin.json                ← Skill manifest\n```\n\n---\n\n## 🎯 Suggested reading paths by role\n\n### I'm new to n8n (never built a workflow)\n1. This file → `01-INSTALL.en.md` → `03-FIRST-WORKFLOW.en.md`\n2. After your first workflow runs, browse `cookbook\u002F00-INDEX.en.md` for your scenario\n3. Stuck? → `04-FAQ.en.md`\n\n### I'm experienced with n8n, evaluating this Pack\n1. This file → `02-USAGE-MODES.en.md`\n2. Read `tests\u002FREPORT-3.md`: real n8n acceptance scores\n3. Browse `examples\u002Ftigerai-flagship\u002F`: enterprise-grade SDD examples\n\n### I'm an engineer \u002F integrator\n1. This file → `spec\u002Fsticky-note-three-layer.md` + `spec\u002Fsticky-note-dsl.md`\n2. `skills\u002Ftigerai\u002Fsticky-note-to-workflow\u002FSKILL.md`: the core executor\n3. `skills\u002Ftigerai\u002Fn8n-api-bridge\u002FSKILL.md`: n8n REST API SOP\n4. `research\u002Fpatterns.md`: 7 standard skeletons + anti-patterns\n\n### I'm distributing this to my team\n1. This file → run `01-INSTALL.en.md` end-to-end\n2. Read `04-FAQ.en.md` to prepare for team questions\n3. Hand the entire folder to teammates and ask them to start at this README\n\n---\n\n## ✨ The three-layer structure (one diagram)\n\n```text\n┌─────────────────────────────────────────────────────┐\n│ 🟡 Layer 1 (yellow sticky): User intent              │\n│    \"Every day at 9 AM...\"                            │\n│    ← AI never modifies this. Always the source of    │\n│      truth.                                          │\n├─────────────────────────────────────────────────────┤\n│    Layer 2: AI-generated nodes & connections        │\n│    Schedule → HTTP → Code → Email                   │\n├─────────────────────────────────────────────────────┤\n│ 🔵 Layer 3 (blue sticky): AI's commentary            │\n│    • Why each node was chosen                        │\n│    • Required credentials                            │\n│    • Assumptions and known limits                    │\n│    • How to test                                     │\n└─────────────────────────────────────────────────────┘\n```\n\n---\n\n## 🛠️ Pain points this Pack solves\n\n| Pain | Solution |\n|---|---|\n| AI-written workflows are inconsistent, hard to review | Enforce three-layer structure |\n| Users don't know how to describe what they want | Plain-language stickies + 8 cookbooks + Q&A mode |\n| AI doesn't know n8n well enough | 7 official Skills + 2,061 workflow corpus |\n| No enterprise-grade patterns | 4 pillars: Atomic Orchestration \u002F Universal Worker \u002F SDD \u002F Security |\n| Don't know where to start | `03-FIRST-WORKFLOW.en.md` 15-min hands-on |\n\n---\n\n## 📊 Real-environment acceptance results (v0.9.0 R3)\n\nTested 8 scenarios on a real n8n 2.10.3 + Postgres setup:\n\n| Layer | Pass rate |\n|---|---|\n| JSON parse | 8\u002F8 (100%) |\n| n8n CLI Import | 8\u002F8 (100%) |\n| API Activate | 7\u002F8 (87.5%) — T3 blocked by real Telegram bot token check |\n| Webhook routing | 4\u002F4 (100%) |\n| Full execute success | 2\u002F4 (with `continueOnFail` design) |\n\nDetails: [`tests\u002FREPORT-3.en.md`](tests\u002FREPORT-3.en.md).\n\n---\n\n## 🔢 Version & changelog\n\nCurrent version: see [`VERSION`](VERSION). All changes: [`CHANGELOG.md`](CHANGELOG.md).\n\n---\n\n## 📜 License\n\n- `skills\u002F_vendor\u002F`: MIT — from [czlonkowski\u002Fn8n-skills](https:\u002F\u002Fgithub.com\u002Fczlonkowski\u002Fn8n-skills), see `skills\u002F_vendor\u002FLICENSE`\n- `reference-workflows\u002F`: MIT — from [Zie619\u002Fn8n-workflows](https:\u002F\u002Fgithub.com\u002FZie619\u002Fn8n-workflows). API tokens, bearer tokens, and other secrets present in the original files have been replaced with placeholders (e.g. `YOUR_API_TOKEN_HERE`) before redistribution.\n- The rest (TigerAI-authored skills, cookbook, specs, docs, install scripts): **TigerAI Proprietary** (distribution terms set by your company)\n\nFull third-party notices: [`THIRD_PARTY_NOTICES.md`](THIRD_PARTY_NOTICES.md).\n\n---\n\n## 🆘 Stuck?\n\nTell Claude \u002F ChatGPT:\n\n> \"I'm new to this. Following the TigerAI Skill Pack README, currently on [filename], hit [problem].\"\n\nThe AI will diagnose. Or check [`04-FAQ.en.md`](04-FAQ.en.md) first.\n","TigerAI n8n Skill Pack 是一个基于自然语言便签构建企业级n8n工作流的AI引擎。它能够将简单的用户意图转化为完整且有文档记录的三层自动化系统，无需编写代码。项目利用了Claude Code和Antigravity等AI代理工具从需求规格到n8n工作流生成的全过程，并包含13个技能、2,061个参考工作流及12种技能模板，支持4种企业模式。特别适合希望探索如何将AI作为工程团队成员使用的开发者或项目经理，以及评估AI代理能否替代手工编写技能\u002F工作流的团队。此外，对于对人机协作产出感兴趣的人来说，该项目也是一个很好的案例研究。",2,"2026-06-06 04:00:41","CREATED_QUERY"]