Tigerai Enterprise Patterns logo

Tigerai Enterprise Patterns

Community
MorrisLu-Taipei
tigerai-enterprise-patterns

Applies TigerAI enterprise-grade design patterns when generating n8n workflows — Atomic Orchestration, Universal Worker (FastAPI), Specification-Driven Development (SDD), and security/governance constraints. Use when the user mentions enterprise / production / 企業級 / 原子化 / orchestration / FastAPI worker / SDD, or when a workflow involves heavy compute (PDF/MP3/image processing), regulated data, or multi-team handoff. Drives architectural decisions like loop transparency (batchSize=1), location-transparent workers, and mandatory error/audit annotations.

Overview

PublisherMorrisLu-Taipei
RepositoryTigerAI-A2A-Code2n8n-Skill-Pack
Skill nametigerai-enterprise-patterns
Stars
86
Forks
26
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by MorrisLu-Taipei on GitHub. Read the source before you install it.

Installation

Install the Tigerai Enterprise Patterns AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/MorrisLu-Taipei/TigerAI-A2A-Code2n8n-Skill-Pack/tree/main/skills/tigerai/tigerai-enterprise-patterns
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/MorrisLu-Taipei/TigerAI-A2A-Code2n8n-Skill-Pack.git /tmp/TigerAI-A2A-Code2n8n-Skill-Pack
mkdir -p .claude/skills
cp -r /tmp/TigerAI-A2A-Code2n8n-Skill-Pack/skills/tigerai/tigerai-enterprise-patterns .claude/skills/tigerai-enterprise-patterns
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Tigerai Enterprise Patterns in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Tigerai Enterprise Patterns on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Tigerai Enterprise Patterns is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

TigerAI Enterprise Patterns — 企業級模式 Skill

🌐 English | 繁體中文

🧱 Drop-in 模板(v0.26.0 起):Retry / Approval / Handover 三個經典模式有可直接 import 的最小工作流 JSON:examples/templates/。AI 套用本 Skill 時可優先參考這三個檔案的節點結構、sticky note 寫法、correlation ID 命名,並修改成使用者場景。

1. 觸發條件

精確觸發詞:

  • 「企業級」/「production」/「正式環境」/「上線」
  • 「原子化」/「atomic」/「orchestration」/「編排」
  • 「FastAPI worker」/「Universal Worker」/「外部 worker」
  • 「SDD」/「specification-driven」/「規格驅動」

隱式觸發: 工作流涉及以下情境,AI 必須主動套用本 Skill:

  • 處理檔案(PDF / MP3 / 圖片 / 文檔)— 套 Universal Worker
  • 多團隊交接 / 跨服務編排 — 套 SDD
  • 法務 / 金融 / 醫療 / PII 資料 — 套安全治理
  • 失敗成本高的任務(金流、ETL)— 套全域 Error Trigger workflow

2. TigerAI 四大支柱(Consultancy Pillars)

源自專案 README.md,AI 產出 workflow 時必須體現:

Pillar 1:原子化編排(Atomic Orchestration)

問題:傳統黑箱 Python 腳本不可見、出錯難 debug。 解法:把流程拆成原子動作,每動作是一個 n8n 節點,loop 用 splitInBatches batchSize=1 讓每 iteration 在 UI 上獨立可見。 AI 規則:

  • 涉及檔案逐筆處理 → splitInBatches batchSize=1(犧牲速度換透明度)
  • 每個原子動作節點的 notes 欄位必填(對應 Layer 1 哪個 step)
  • 不允許 1 個 code 節點塞 100 行邏輯 → 拆成多節點

Pillar 2:Universal Worker(位置透明)

問題:n8n 容器內裝 FFmpeg/PyMuPDF 會污染環境。 解法:把重邏輯封裝成 FastAPI Worker 容器,n8n 只用 httpRequest 呼叫。Worker 在本機/WSL/雲端皆同樣 URL。 AI 規則:

  • 偵測到「PDF 切分 / MP3 處理 / 圖片轉檔 / OCR / FFmpeg」字眼 → 不要產 code 節點,產 httpRequest 呼叫 worker
  • 預設 worker URL:http://worker:8000/<endpoint>,timeout 300 秒
  • Layer 3 必註明「需先部署 FastAPI Worker container」+ 預期 API 介面

Pillar 3:Skill-Driven Development(本 Pack 自身)

問題:AI 寫的 workflow 破碎、不一致。 解法:以 SKILL.md 為唯一準繩。生產 workflow 前必先讀 spec / cookbook / patterns。 AI 規則:

  • 不憑空產生 — 先比對本 Pack 的 7 大 pattern
  • 不私下修改 spec — 任何疑義由使用者裁決

Pillar 4:企業安全評估

AI 規則:

  • Webhook 預設未驗證 secret → Layer 3 必標「資安風險」
  • 任何 credential 寫死 → 拒絕,要求改 credential reference
  • 涉及 PII 流程 → Layer 3 必標「需資料留存政策確認」
  • timeout 預設值:HTTP 30s / Worker 300s / Wait 不超過 1 hour

Pillar 4.1:時效性警告(Sunset / Deprecation Watch)

外部服務 / API / 節點會被棄用。AI 產出 workflow 時必須主動告知時效風險,避免使用者部署後幾個月內失效。

已知時效性風險清單(持續更新):

服務 / 節點狀態行動建議
LINE Notify API2026/03/31 停用改用 LINE Messaging API(channel access token)
Twitter API v1.12023 已停用,免費 v2 受限評估 X Premium 或改通訊管道
n8n-nodes-base.cron已過時但仍運作改用 n8n-nodes-base.scheduleTrigger
n8n-nodes-base.function / functionItem已過時改用 n8n-nodes-base.code
OpenAI gpt-3.5-turbo模型棄用週期較快預留改 model 的能力(用 expression 帶模型 ID)
Slack Legacy Tokens已停用必用 OAuth2

AI 產出規則:

  • 偵測到「LINE Notify」字眼 → Layer 3 必含「⚠️ 2026/03/31 停用警告」
  • 偵測到 cron / function / functionItem 節點 → 拒絕產出,改用新節點
  • 偵測到 LLM 模型字眼 → Layer 3 加「模型可能 18 個月內棄用,建議 model ID 用 env var 帶入」
  • 不在清單內但使用者明示「已知會停用」→ 同步轉錄至 Layer 3

追蹤責任:

  • 本清單由 TigerAI 維運團隊至少每季更新一次
  • 客戶可透過 Slack channel #tigerai-skill-pack-updates 訂閱新增項目
  • 過時項目移至 research/deprecated-services.md(暫未建立)保留歷史

Pillar 4.2:人工核准 / Handover 設計(Human-in-the-Loop)

問題:AI 寫的 workflow 常常「全自動跑完」,但企業實際需要在關鍵節點停下來等真人簽核(出帳前、發大量訊息前、修改正式資料前)。沒有 approval 節點=出事沒人擋。

核准節點型態(n8n 內建):

通道節點適用情境
Emailn8n-nodes-base.emailSend + wait + 回覆 webhook,或 sendAndWaitForResponse 模式主管慢但留紀錄
Slackslack send + wait for button click webhook即時、多人會議室可見
Form / 內建 Form TriggerformTrigger 收第二階段表單跨部門、外部廠商
Telegram / LINEhttpRequest 發 inline keyboard一線值班
n8n built-in sendAndWait operationEmail / Slack / Teams 都支援首選,內建超時與簽到

設計規則(AI 必遵守):

  1. 每個 approval 節點 timeout 必設。預設 24 小時;金流類 4 小時;客服類 2 小時。
  2. Timeout 後必有 escalation 路徑。三種模式(依風險升序):
    • 自動拒絕並通知申請者(最保守)
    • 轉給代理人或主管的主管(升級)
    • 自動核准但標「待事後追認」(最寬鬆,僅限低風險)
  3. Audit trail 強制留:在 approval 節點之後接一個 audit log 寫入(Sheets / Postgres / Slack channel 都可),欄位至少:request_id, requester, approver, decision, timestamp, reason, ip_or_channel。
  4. 拒絕路徑必寫補償動作:「IF 拒絕 → 回滾 / 通知 / 寫拒絕原因到 log」。不能只是流程結束。
  5. 不可把 approval 節點接的 token 寫進 Layer 1 sticky note — token 是執行期參數,屬 credential。

Handover(移交)設計規則:

場景必做
真人客服接手(LINE / Web chat)寫 is_human_mode 狀態到 DB;timeout 後自動歸還;Layer 3 註明逾時策略
跨班次交班寫一個 daily-handover-summary workflow(schedule + 查當天未完成任務 + 發 Slack)
AI → 工程師人工接管工作流暫停(wait until webhook),把 context 寫成可閱讀的 Markdown 摘要存到 audit log

反模式(拒絕產出):

  • ❌ Approval 節點無 timeout
  • ❌ Timeout 後直接「continue」當作通過
  • ❌ Approval 結果沒寫進任何 log
  • ❌ 拒絕路徑跟通過路徑共用節點且沒分支
  • ❌ 在 Layer 1 sticky 寫死 approver email(應走 credential 或 settings 表)

對應 onprem LINE CS 案例:那個系統的「真人接手」流程是 handover 半套示範(有 is_human_mode 狀態 + 逾時歸還)。但 SECURITY-CAVEATS 也指出後台沒有真實 auth — 所以 approval 本身的「按下去那個人是誰」這層保證還沒做完。AI 產出企業級 workflow 時必確認 approver 身分驗證真實存在而不是假樁。


3. SDD(Specification-Driven Development)整合

對企業級 workflow,AI 在產出 JSON 同時,建議使用者另外建立 specification.md 記錄:

markdown
# Workflow Specification

## 1. Purpose
- Business objective
- Stakeholders

## 2. Trigger & Inputs
- Trigger type
- Input schema (JSON Schema 或欄位清單)

## 3. Business Logic (Step-by-Step)
- 對應本 workflow 的 @step 順序

## 4. Outputs
- 對外回傳 / 寫入目標 / 通知對象

## 5. Errors & Recovery
- 已知失敗模式
- Error Trigger workflow ID

## 6. Test Scenarios
- 至少 3 case:golden path / edge / error

AI 規則:當使用者明示「企業級」/「SDD」時,產出 JSON 後主動詢問:「要不要我同步產生 specification.md?」


4. 旗艦範例對照

範例體現 Pillar路徑
splitPDF-orchestrated1 + 2examples/tigerai-flagship/splitPDF-orchestrated/
splitMP3-API-Orchestrated1 + 2examples/tigerai-flagship/splitMP3-API-Orchestrated/
openwebui-bridge-v23(Skill driven)examples/tigerai-flagship/openwebui-bridge-v2/
Test 4 (本 Pack)全部 4tests/4-pdf-worker-s3/workflow.json

5. AI 自我檢查清單

產出企業級 workflow 後,必逐條自查:

  • 重邏輯放 worker,n8n 只編排?
  • loop 是否 batchSize=1(除非明示要速度)?
  • 每節點 notes 都填?
  • credential 都是 reference(不是寫死)?
  • Layer 3 含「資安風險」/「資料留存」段?
  • 全域 Error Trigger workflow 已建議?
  • timeout 都設?
  • 對應 SDD 的 6 段是否都能在 Layer 1+3 找到?

少一條即不算企業級。


6. 與其他 Skill 整合

  • 與 sticky-note-to-workflow:本 Skill 是其「企業級分支」,當偵測企業情境自動啟用
  • 與 n8n-validation-expert:產出後先過 validation profile=strict

7. 反模式(企業情境下絕對禁止)

  • ❌ Code 節點塞 100 行業務邏輯
  • ❌ 寫死 API key / DB password
  • ❌ 用 manualTrigger 上線
  • ❌ 沒設 timeout 直接呼叫第三方 API
  • ❌ 沒有 Error Trigger workflow
  • ❌ 對 PII 資料未做存取控管 / log redaction
  • ❌ Workflow 名稱用「Untitled」/「Test」/「Copy of...」

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Tigerai Enterprise Patterns AI skill do?

Applies TigerAI enterprise-grade design patterns when generating n8n workflows — Atomic Orchestration, Universal Worker (FastAPI), Specification-Driven Development (SDD), and security/governance constraints. Use when the user mentions enterprise / production / 企業級 / 原子化 / orchestration / FastAPI worker / SDD, or when a workflow involves heavy compute (PDF/MP3/image processing), regulated data, or multi-team handoff. Drives architectural decisions like loop transparency (batchSize=1), location-transparent workers, and mandatory error/audit annotations.

Why use Tigerai Enterprise Patterns on TypingMind?

Because you install it once and use it with any model. Tigerai Enterprise Patterns is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Tigerai Enterprise Patterns in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/MorrisLu-Taipei/TigerAI-A2A-Code2n8n-Skill-Pack/tree/main/skills/tigerai/tigerai-enterprise-patterns. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Tigerai Enterprise Patterns?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Tigerai Enterprise Patterns?

As many as you like. As long as a model supports skills, you can use Tigerai Enterprise Patterns with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Tigerai Enterprise Patterns AI skill free?

Yes. It is published on GitHub by MorrisLu-Taipei under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇