slack-to-teams

Microsoft ｜ Agent 與工作流 ｜ 原文 ｜ MIT

把 Slack bot 遷移到 Teams、跨平台橋接的路由式專家系統

---
name: slack-to-teams
description: "microsoft-365-agents-toolkit 的子 skill。這是一套有 100 多個微型專家檔案的路由式專家系統，用於把 Slack bot 遷移到 Teams、跨平台橋接，以及雙平台 bot 開發。USE FOR: migrating Slack bot to Teams、adding Teams support to Slack bot、building dual-platform bots、converting Block Kit to Adaptive Cards、identity/OAuth bridging、deploying bots to Azure or AWS、configuring AI model providers。DO NOT USE FOR: 一般網頁開發、非 bot 專案、沒有 Slack 的獨立 Teams 開發（請改用上層 skill）。"
---

# Slack to Teams 專家系統

一套有 100 多個微型專家檔案的路由式專家系統，用於把 Slack bot 遷移到 Teams，以及建立跨平台 bot。

> **上層 skill：** ATK CLI 的設定和路由請看 [../SKILL.md](../SKILL.md)。本機測試請看 [../test-playground/test-playground.md](../test-playground/test-playground.md) 或 [../test-teams/test-teams.md](../test-teams/test-teams.md)。雲端 deploy 請看 [../provision-deploy/provision-deploy.md](../provision-deploy/provision-deploy.md)。疑難排解請看 [../troubleshoot/troubleshoot.md](../troubleshoot/troubleshoot.md)。

## 什麼時候用

- 建立新的 Slack bot、Teams bot 或雙平台 bot
- 為現有的 Slack bot 加上 Teams 支援（或反過來）
- 在平台之間遷移 bot
- 把 bot deploy 到 Azure 或 AWS
- 為 bot 設定 AI 模型供應商
- 在 Block Kit 和 Adaptive Cards 之間轉換 UI
- 橋接平台之間的身分、事件、檔案或傳輸層
- 讓 Teams bot 專案相容於 Microsoft Agents Toolkit（m365agents.yml、env/、appPackage 的 placeholder）

## 步驟

### 步驟 1：判斷專案類型

評估開發者是已經有程式碼，還是從零開始。

- **新專案** → 前往[新專案流程](#new-project-flow)
- **現有專案** → 前往[現有專案流程](#existing-project-flow)

---

### 新專案流程

#### 1a：選擇平台

問開發者要支援哪些平台：
- Slack + Teams（雙平台）
- 只要 Slack
- 只要 Teams

#### 1b：載入專家系統

1. 讀 [experts/index.md](../experts/index.md)，這是根路由器。
2. 執行該檔案裡定義的**事前訪談**。
3. 依平台選擇路由：
   - **Slack + Teams** → 讀 [experts/bridge/index.md](../experts/bridge/index.md)。載入 [cross-platform-advisor](../experts/bridge/cross-platform-advisor-ts.md)。
   - **只要 Slack** → 讀 [experts/slack/index.md](../experts/slack/index.md)。
   - **只要 Teams** → 讀 [experts/teams/index.md](../experts/teams/index.md)。

#### 1c：架構設定

1. 讀 [cross-platform-architecture](../experts/bridge/cross-platform-architecture-ts.md)（就算只做單一平台也要讀，它會建立之後加第二個平台的模式）。
2. 交給 domain router 和 advisor 接手。
3. 在目標專案根目錄寫一份 `PLAN.md`，記錄平台、功能、載入的專家。

#### 1d：實作

照 advisor 或 domain router 的輸出做。一個功能一個功能實作：
1. 從有優先順序的清單挑下一個功能。
2. 載入該功能指定的專家。
3. 用專家的模式和規則來實作。
4. 對照專家的陷阱段落做驗證。

---

### 現有專案流程

#### 2a：分析專案

**平行**跑這四個子分析：

**偵測語言**：掃描 `package.json`+`tsconfig.json`（TypeScript）、`pom.xml`/`build.gradle`（Java）、`*.csproj`/`*.sln`（C#）、`go.mod`（Go）、`requirements.txt`/`pyproject.toml`（Python）、`Gemfile`（Ruby）、`Cargo.toml`（Rust）。

**偵測目前的平台**：掃描相依套件裡的 SDK 指標：
- Slack：`@slack/bolt`、`@slack/web-api`、`slack_bolt`、`app.message`、`app.command`、`ack()`
- Teams：`@microsoft/teams-ai`、`botbuilder`、`TeamsActivityHandler`、`app.turn`、Adaptive Cards

**偵測功能**：掃描 slash command、Block Kit/Adaptive Cards、action handler、OAuth、檔案上傳/下載、排程、threading、AI/LLM 呼叫、主動訊息。

**偵測架構**：掃描 web framework（Express、Fastify 等）、hosting 目標（Azure、AWS、Docker）、雲端供應商、架構模式（單一 bot、雙 bot、monolith）。

#### 2b：語言關卡

把偵測到的語言分類到 SDK 等級：

| 等級 | 語言 | 指引 |
|---|---|---|
| **1：完整 SDK** | TypeScript / JavaScript | 有完整的專家系統可用 |
| **2：改寫** | Python | 兩邊的 SDK 都有，改寫 TS 的模式。載入 [bolt-python](../experts/slack/bolt-python.md)、[teams-python](../experts/teams/teams-python.md)、[python-cross-platform](../experts/bridge/python-cross-platform.md) |
| **3：SDK 分裂** | Java、C# | 一個平台有 SDK，另一個要走 REST。載入 [bolt-java](../experts/slack/bolt-java.md) 或 [teams-dotnet](../experts/teams/teams-dotnet.md) + [rest-only](../experts/bridge/rest-only-integration-ts.md) |
| **4：沒有 SDK** | Go、Ruby、Rust | 兩邊都只能用 REST。載入 [rest-only](../experts/bridge/rest-only-integration-ts.md) |

#### 2c：專家覆蓋缺口分析

1. 讀 [experts/analyzer.md](../experts/analyzer.md)。
2. 執行 analyzer 的流程：掃描 manifest、相依套件、原始碼。
3. 把偵測到的技術和現有專家交叉比對。
4. 提出缺口分析：哪些技術有覆蓋、哪些沒有。

#### 2d：建立缺少的專家（需要時）

1. 讀 [experts/builder.md](../experts/builder.md)。
2. 對每個缺口：分析專案的使用方式 → 讀套件原始碼 → 起草專家 → 驗證 → 接進路由。
3. 用 [experts/_expert-ts.md](../experts/_expert-ts.md) 當範本。

#### 2e：載入專家系統

1. 讀 [experts/index.md](../experts/index.md)，這是根路由器。
2. 執行事前訪談，用 2a 的分析結果預先填答案。
3. 依偵測到的平台和任務意圖路由：
   - 有 Slack、想要 Teams → **Bridge** domain
   - 有 Teams、想要 Slack → **Bridge** domain
   - 兩個都有 → 依任務路由（橋接調整、deploy、模型等）
   - 兩個都沒有 → 問要以哪些平台為目標
4. 如果是做橋接，載入 [cross-platform-advisor](../experts/bridge/cross-platform-advisor-ts.md)。把分析結果餵進 Phase 1。

#### 2f：架構與實作

1. 讀 [cross-platform-architecture](../experts/bridge/cross-platform-architecture-ts.md)。
2. 寫一份 `PLAN.md`，記錄分析結果、路由決定和功能遷移順序。
3. 依 advisor 有優先順序的清單，一個功能一個功能實作。

## Agents Toolkit 相容性

ATK 相容的專案結構（m365agents.yml、env/ 檔案、appPackage）請看上層 skill：
- **[上層 SKILL.md](../SKILL.md)**：ATK CLI 設定、子 skill 路由、流程鏈
- **[manifest-and-yaml.md](../toolkit/manifest-and-yaml.md)**：專案檔案、YAML 設定、環境變數、.localConfigs 流程
- **[commands.md](../toolkit/commands.md)**：打包、驗證、分享、協作、環境管理
- **[templates.md](../toolkit/templates.md)**：所有範本及其語言支援

Teams manifest schema 和打包請看 [runtime.manifest-ts](../experts/teams/runtime.manifest-ts.md)。

## 錯誤復原

如果專家系統沒涵蓋某個主題：
1. 讀 [experts/fallback.md](../experts/fallback.md)。
2. Phase 1：重新掃描所有 domain router，找漏掉的專家。
3. Phase 2：用網路搜尋補剩下的知識缺口。
4. 考慮用 [experts/builder.md](../experts/builder.md) 建立新專家。

## 專家 Domain

| Domain | 索引 | 說明 |
|---|---|---|
| Slack | [experts/slack/index.md](../experts/slack/index.md) | Bolt framework、事件、OAuth、指令、UI、CLI |
| Teams | [experts/teams/index.md](../experts/teams/index.md) | Teams AI SDK、Adaptive Cards、Graph、MCP、A2A、deploy |
| Bridge | [experts/bridge/index.md](../experts/bridge/index.md) | 27 個跨平台轉換專家（核心差異化所在） |
| Deploy | [experts/deploy/index.md](../experts/deploy/index.md) | Azure 與 AWS 的 deploy 逐步教學 |
| Models | [experts/models/index.md](../experts/models/index.md) | AI 模型供應商（OpenAI、Anthropic、Bedrock 等） |
| Convert | [experts/convert/index.md](../experts/convert/index.md) | 轉換成 TypeScript 的語言轉換 |
| Security | [experts/security/index.md](../experts/security/index.md) | 輸入驗證、secrets 管理 |

## 平台比較文件

並排比較兩個平台的參考指南：

- [UI Components](../docs/ui-components.md)：Block Kit 對 Adaptive Cards
- [Messaging & Commands](../docs/messaging-and-commands.md)
- [Identity & Auth](../docs/identity-and-auth.md)
- [Interactive Responses](../docs/interactive-responses.md)
- [Files & Links](../docs/files-and-links.md)
- [Middleware & Handlers](../docs/middleware-and-handlers.md)
- [Infrastructure](../docs/infrastructure.md)
- [Advanced Features](../docs/advanced-features.md)
- [Feature Gaps](../docs/feature-gaps.md)
- [Workflow Scenarios](../docs/workflows.md)：給 Teams bot 用的訊息原生工作流程模式（觸發、狀態、邏輯、AI、可見度）

---
name: slack-to-teams
description: "Sub-skill of microsoft-365-agents-toolkit. Routed expert system with 100+ micro-expert files for migrating Slack bots to Teams, cross-platform bridging, and dual-platform bot development. USE FOR: migrating Slack bot to Teams, adding Teams support to Slack bot, building dual-platform bots, converting Block Kit to Adaptive Cards, identity/OAuth bridging, deploying bots to Azure or AWS, configuring AI model providers. DO NOT USE FOR: general web development, non-bot projects, standalone Teams development without Slack (use parent skill instead)."
---

# Slack to Teams Expert System

A routed expert system with 100+ micro-expert files for migrating Slack bots to Teams and building cross-platform bots.

> **Parent skill:** For ATK CLI setup and routing, see [../SKILL.md](../SKILL.md). For local testing, see [../test-playground/test-playground.md](../test-playground/test-playground.md) or [../test-teams/test-teams.md](../test-teams/test-teams.md). For cloud deploy, see [../provision-deploy/provision-deploy.md](../provision-deploy/provision-deploy.md). For troubleshooting, see [../troubleshoot/troubleshoot.md](../troubleshoot/troubleshoot.md).

## When to Use

- Building a new Slack bot, Teams bot, or dual-platform bot
- Adding Teams support to an existing Slack bot (or vice versa)
- Migrating a bot between platforms
- Deploying a bot to Azure or AWS
- Configuring AI model providers for a bot
- Converting UI between Block Kit and Adaptive Cards
- Bridging identity, events, files, or transport between platforms
- Making a Teams bot project compatible with Microsoft Agents Toolkit (m365agents.yml, env/, appPackage placeholders)

## Procedure

### Step 1: Determine project type

Assess whether the developer has an existing codebase or is starting fresh.

- **New project** → go to [New Project Flow](#new-project-flow)
- **Existing project** → go to [Existing Project Flow](#existing-project-flow)

---

### New Project Flow

#### 1a: Platform selection

Ask the developer which platform(s) to support:
- Slack + Teams (dual-platform)
- Slack only
- Teams only

#### 1b: Load the expert system

1. Read [experts/index.md](../experts/index.md) — the root router.
2. Execute the **pre-task interview** defined in that file.
3. Route based on platform choice:
   - **Slack + Teams** → Read [experts/bridge/index.md](../experts/bridge/index.md). Load [cross-platform-advisor](../experts/bridge/cross-platform-advisor-ts.md).
   - **Slack only** → Read [experts/slack/index.md](../experts/slack/index.md).
   - **Teams only** → Read [experts/teams/index.md](../experts/teams/index.md).

#### 1c: Architecture setup

1. Read [cross-platform-architecture](../experts/bridge/cross-platform-architecture-ts.md) (even for single-platform — establishes patterns for adding a second platform later).
2. Let the domain router and advisor take over.
3. Write a `PLAN.md` in the target project root with platform, features, experts loaded.

#### 1d: Implementation

Follow the advisor's or domain router's output. Implement feature by feature:
1. Pick the next feature from the prioritized list.
2. Load the expert(s) specified for that feature.
3. Implement using the expert's patterns and rules.
4. Verify against the expert's pitfalls section.

---

### Existing Project Flow

#### 2a: Analyze the project

Run these four sub-analyses **in parallel**:

**Detect language** — Scan for `package.json`+`tsconfig.json` (TypeScript), `pom.xml`/`build.gradle` (Java), `*.csproj`/`*.sln` (C#), `go.mod` (Go), `requirements.txt`/`pyproject.toml` (Python), `Gemfile` (Ruby), `Cargo.toml` (Rust).

**Detect current platform** — Scan dependencies for SDK indicators:
- Slack: `@slack/bolt`, `@slack/web-api`, `slack_bolt`, `app.message`, `app.command`, `ack()`
- Teams: `@microsoft/teams-ai`, `botbuilder`, `TeamsActivityHandler`, `app.turn`, Adaptive Cards

**Detect features** — Scan for slash commands, Block Kit/Adaptive Cards, action handlers, OAuth, file upload/download, scheduling, threading, AI/LLM calls, proactive messages.

**Detect architecture** — Scan for web framework (Express, Fastify, etc.), hosting target (Azure, AWS, Docker), cloud provider, architecture pattern (single bot, dual-bot, monolith).

#### 2b: Language gate

Classify the detected language into SDK tiers:

| Tier | Languages | Guidance |
|---|---|---|
| **1: Full SDK** | TypeScript / JavaScript | Full expert system available |
| **2: Adapt** | Python | Both SDKs exist — adapt TS patterns. Load [bolt-python](../experts/slack/bolt-python.md), [teams-python](../experts/teams/teams-python.md), [python-cross-platform](../experts/bridge/python-cross-platform.md) |
| **3: Split SDK** | Java, C# | One platform has SDK, other needs REST. Load [bolt-java](../experts/slack/bolt-java.md) or [teams-dotnet](../experts/teams/teams-dotnet.md) + [rest-only](../experts/bridge/rest-only-integration-ts.md) |
| **4: No SDK** | Go, Ruby, Rust | REST-only for both. Load [rest-only](../experts/bridge/rest-only-integration-ts.md) |

#### 2c: Expert coverage gap analysis

1. Read [experts/analyzer.md](../experts/analyzer.md).
2. Execute the analyzer workflow: scan manifests, dependencies, source files.
3. Cross-reference detected tech against existing experts.
4. Present gap analysis — covered vs uncovered technologies.

#### 2d: Build missing experts (if needed)

1. Read [experts/builder.md](../experts/builder.md).
2. For each gap: analyze project usage → read package source → draft expert → validate → wire into routing.
3. Use [experts/_expert-ts.md](../experts/_expert-ts.md) as the template.

#### 2e: Load the expert system

1. Read [experts/index.md](../experts/index.md) — the root router.
2. Execute the pre-task interview, pre-filling with analysis results from 2a.
3. Route based on detected platform and task intent:
   - Has Slack, wants Teams → **Bridge** domain
   - Has Teams, wants Slack → **Bridge** domain
   - Has both → Route by task (bridge refinement, deploy, models, etc.)
   - Has neither → Ask which platform(s) to target
4. Load [cross-platform-advisor](../experts/bridge/cross-platform-advisor-ts.md) if bridging. Feed analysis results into Phase 1.

#### 2f: Architecture and implementation

1. Read [cross-platform-architecture](../experts/bridge/cross-platform-architecture-ts.md).
2. Write a `PLAN.md` with analysis results, routing decisions, and feature migration order.
3. Implement feature by feature using the advisor's prioritized list.

## Agents Toolkit Compatibility

For ATK-compatible project structure (m365agents.yml, env/ files, appPackage), see the parent skill:
- **[Parent SKILL.md](../SKILL.md)** — ATK CLI setup, sub-skill routing, workflow chains
- **[manifest-and-yaml.md](../toolkit/manifest-and-yaml.md)** — Project files, YAML config, env vars, .localConfigs flow
- **[commands.md](../toolkit/commands.md)** — Package, validate, share, collaborate, environment management
- **[templates.md](../toolkit/templates.md)** — All templates with language support

For Teams manifest schema and packaging, see [runtime.manifest-ts](../experts/teams/runtime.manifest-ts.md).

## Error Recovery

If the expert system fails to cover a topic:
1. Read [experts/fallback.md](../experts/fallback.md).
2. Phase 1: Re-scan all domain routers for missed experts.
3. Phase 2: Web-search for remaining knowledge gaps.
4. Consider creating a new expert using [experts/builder.md](../experts/builder.md).

## Expert Domains

| Domain | Index | Description |
|---|---|---|
| Slack | [experts/slack/index.md](../experts/slack/index.md) | Bolt framework, events, OAuth, commands, UI, CLI |
| Teams | [experts/teams/index.md](../experts/teams/index.md) | Teams AI SDK, Adaptive Cards, Graph, MCP, A2A, deploy |
| Bridge | [experts/bridge/index.md](../experts/bridge/index.md) | 27 cross-platform conversion experts (the core differentiator) |
| Deploy | [experts/deploy/index.md](../experts/deploy/index.md) | Azure & AWS deployment walkthroughs |
| Models | [experts/models/index.md](../experts/models/index.md) | AI model providers (OpenAI, Anthropic, Bedrock, etc.) |
| Convert | [experts/convert/index.md](../experts/convert/index.md) | Language conversion to TypeScript |
| Security | [experts/security/index.md](../experts/security/index.md) | Input validation, secrets management |

## Platform Comparison Docs

Reference guides for side-by-side platform comparison:

- [UI Components](../docs/ui-components.md) — Block Kit vs Adaptive Cards
- [Messaging & Commands](../docs/messaging-and-commands.md)
- [Identity & Auth](../docs/identity-and-auth.md)
- [Interactive Responses](../docs/interactive-responses.md)
- [Files & Links](../docs/files-and-links.md)
- [Middleware & Handlers](../docs/middleware-and-handlers.md)
- [Infrastructure](../docs/infrastructure.md)
- [Advanced Features](../docs/advanced-features.md)
- [Feature Gaps](../docs/feature-gaps.md)
- [Workflow Scenarios](../docs/workflows.md) — Message-native workflow patterns (triggers, state, logic, AI, visibility) for Teams bots