PM-Kit
Integrates AI assistants with product management tools like Linear and Slack, providing specialized agents for discovery, strategy, prioritization, roadmap planning, and metrics definition via the Model Context Protocol.
README
PM-Kit
Product Management AI Toolkit - A suite of specialized AI agents and integration tools for modern Product Managers.
PM-Kit connects your AI (Claude Desktop) with your product management stack (Linear, Slack, Google Drive) using the Model Context Protocol (MCP).
🚀 Active Tools
🛠️ Core PM Agents
These agents use a built-in knowledge base of product management best practices.
| Tool Name | Agent Role | Use Case |
|---|---|---|
pm_discover |
Discovery | User research, problem analysis, interviews |
pm_strategy |
Strategy | PRDs, vision docs, risk assessment |
pm_prioritize |
Prioritization | RICE scoring, backlog refinement |
pm_roadmap |
Roadmap | Milestone planning, release timelines |
pm_mvp |
MVP Scoping | Scoping minimum viable features |
pm_metrics |
Metrics | Defining KPIs and success criteria |
pm_consult |
Handbook | Consult the handbook to pick the right agent |
📋 Agent Cheat Sheet (Natural Language)
You can trigger these agents in Claude using natural language.
| Agent | Command Trigger | Use Case |
|---|---|---|
| Product Discovery | "Discover...", "Research..." | Analyzing user problems, interviews |
| Strategy | "Create strategy...", "Vision..." | Drafting PRDs, strategy docs |
| Prioritization | "Prioritize...", "Rank backlog..." | RICE scoring, ordering tasks |
| Roadmap | "Create roadmap...", "Plan Q1..." | Timeline planning |
| MVP Scoping | "Define MVP...", "Scope..." | Defining minimum viable features |
| Metrics | "Define metrics...", "KPIs..." | Success criteria |
🔌 Integration Tools
Direct actions for your workspace.
| Tool Name | Integration | Description |
|---|---|---|
linear_create_issue |
Linear | Create tickets/issues directly |
linear_update_issue |
Linear | Update title, description, or priority |
linear_delete_issue |
Linear | Permanently remove an issue |
linear_get_projects |
Linear | List all active projects |
linear_create_project |
Linear | Start a new project with description |
linear_update_project |
Linear | Update project name or description |
linear_delete_project |
Linear | Archive a project |
linear_get_documents |
Linear | List all workspace documents |
slack_send_message |
Slack | Post messages to any channel |
slack_get_channels |
Slack | List available public channels |
🛠️ Setup Guide (Local Mode)
The most stable way to run PM-Kit is locally on your machine.
1. Requirements
- Node.js: v18+ (verified on v22)
- Claude Desktop
2. Integration API Keys
Linear API Key
- Go to linear.app/settings → Security & access
- Create a "Personal API key".
Slack Bot Token
- Go to api.slack.com/apps → Create an App → From Scratch.
- In "OAuth & Permissions", add
chat:writeandchannels:readscopes. - Install to workspace and copy the "Bot User OAuth Token" (
xoxb-).
3. Claude Desktop Configuration
Open your Claude Desktop config file:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - Mac:
~/Library/Application Support/Claude/claude_desktop_config.json
Add the following (update the path to match your local installation):
{
"mcpServers": {
"pm-kit": {
"command": "c:/Visual Code Projects/PM-kit/run_local.bat",
"args": [],
"env": {
"LINEAR_KEY": "lin_api_YOUR_LINEAR_KEY",
"SLACK_TOKEN": "xoxb-YOUR_SLACK_TOKEN"
}
}
}
}
4. Apply Changes
- Quit Claude Desktop (Right-click in system tray → Quit).
- Restart Claude Desktop.
- You should now see the
pm-kittools in the bottom bar!
🏗️ Development
Local Build
npm install
npm run build
Test Scripts
# Verify connection to Linear/Slack
npm run test:integrations
📄 License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。