teamdynamix-mcp
Enables AI agents to interact with TeamDynamix ITSM operations including tickets, assets, KB, and more through safety-gated tools.
README
🎟️ TeamDynamix MCP Server (unofficial)
A TypeScript Model Context Protocol (MCP) server that exposes TeamDynamix ITSM capabilities as agent-callable tools. Designed for AI agents and MCP clients that need structured, safety-gated access to TeamDynamix operations.
Features
- Powered by the official TeamDynamix TypeScript SDK: type-safe, auto-generated API methods with built-in retry, exponential backoff, runtime validation, and token lifecycle management
- 11 domain gateway tools that route validated actions across discovery, tickets, relationships, KB, assets, CMDB, people, services, projects, time, and reference data
- Safe by default: write tools disabled until explicitly opted in;
destructive operations require
confirm: true - Two auth modes: standard (username/password) and admin (BEID/WebServicesKey)
- Covers 207 API endpoints across all TeamDynamix domains, verified against the SDK route manifest
- Rate-limit aware: SDK built-in retry with exponential backoff on 429 and 5xx responses
- Zod-validated inputs: schema enforcement before any API call
- Agent skill and prompt included: ready-to-use skill definition for GitHub Copilot and compatible agents
MCP Registry
This server is published to the MCP Registry as io.github.selfagency/teamdynamix-mcp.
- Registry name:
io.github.selfagency/teamdynamix-mcp - Install via registry:
mcp-install io.github.selfagency/teamdynamix-mcp(when available) - Manual install: Use npm package
@selfagency/teamdynamix-mcpas shown below
Quick start (developer setup)
This quick start is for contributors running the server from source in this
repository. If you are a regular MCP client user, skip to
MCP client configuration and use the npx
command examples.
# 1. Clone the repository
git clone https://github.com/selfagency/teamdynamix-mcp.git
cd teamdynamix-mcp
# 2. Install dependencies
pnpm install
# 3. Configure environment
cp .env.example .env
# Edit .env with your TeamDynamix credentials and base URL
# 4. Start the development server
pnpm dev
Minimum required environment variables
TEAMDYNAMIX_BASE_URL=https://your-tenant.teamdynamix.com/TDWebApi
TEAMDYNAMIX_AUTH_MODE=standard # or: admin
TEAMDYNAMIX_USERNAME=you@example.com # standard mode
TEAMDYNAMIX_PASSWORD=your-password # standard mode
For admin mode, use TEAMDYNAMIX_BEID and TEAMDYNAMIX_WEB_SERVICES_KEY instead.
MCP client configuration
Security notice: Values shown above are placeholders. Never commit real credentials to source control. Inject secrets at runtime via a
.envfile (excluded by.gitignore), CI/CD secrets, or a secret manager.
VS Code (.vscode/mcp.json)
{
"servers": {
"teamdynamix": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@selfagency/teamdynamix-mcp"],
"env": {
"TEAMDYNAMIX_BASE_URL": "https://your-tenant.teamdynamix.com/TDWebApi",
"TEAMDYNAMIX_AUTH_MODE": "standard",
"TEAMDYNAMIX_USERNAME": "you@example.com",
"TEAMDYNAMIX_PASSWORD": "your-password"
}
}
}
}
Claude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"teamdynamix": {
"command": "npx",
"args": ["-y", "@selfagency/teamdynamix-mcp"],
"env": {
"TEAMDYNAMIX_BASE_URL": "https://your-tenant.teamdynamix.com/TDWebApi",
"TEAMDYNAMIX_AUTH_MODE": "standard",
"TEAMDYNAMIX_USERNAME": "you@example.com",
"TEAMDYNAMIX_PASSWORD": "your-password"
}
}
}
}
Production (built bundle)
pnpm build # outputs to dist/
{
"command": "npx",
"args": ["-y", "@selfagency/teamdynamix-mcp"]
}
Verify connectivity
After connecting your MCP client, call teamdynamix_discovery with:
action: "server_status"payload: {}response_format: "json"
A successful response shows status.configured: true.
Domain gateway tools
Each gateway tool accepts:
action: domain-specific operation namepayload: object for that action’s parametersresponse_format:"markdown"or"json"
| Domain | Gateway tool |
|---|---|
| Discovery | teamdynamix_discovery |
| Tickets | teamdynamix_tickets |
| Ticket relationships | teamdynamix_ticket_relationships |
| Knowledge Base | teamdynamix_knowledge_base |
| Assets | teamdynamix_assets |
| CMDB | teamdynamix_cmdb |
| People | teamdynamix_people |
| Services | teamdynamix_services |
| Projects | teamdynamix_projects |
| Time | teamdynamix_time |
| Reference data | teamdynamix_reference_data |
† Requires confirm: true in addition to write tools being enabled.
All write/mutating actions require TEAMDYNAMIX_ENABLE_WRITE_TOOLS=true.
Safety defaults
| Flag | Default | Effect when true |
|---|---|---|
TEAMDYNAMIX_ENABLE_WRITE_TOOLS |
false |
Enables all create/update/comment/mutation tools |
TEAMDYNAMIX_ENABLE_ADMIN_TOOLS |
false |
Enables admin-scope operations |
Destructive unlink operations additionally require confirm: true in the
tool call regardless of write flag state.
Agent skill
This repository ships a GitHub Copilot-compatible skill definition:
- Skill:
skills/teamdynamix/SKILL.md— ID-first workflow patterns, safety branching rules, runbooks, gotchas, and tool catalog - Agent prompt:
prompts/teamdynamix-agent.prompt.md— system prompt for agent mode with TeamDynamix context
Install the skill via your Copilot skill configuration
(see skills/teamdynamix/SKILL.md for full details).
Documentation
| Section | Purpose |
|---|---|
| Tutorials | Step-by-step setup and first workflows |
| How-to guides | Task-oriented recipes for real operations |
| Reference | Exhaustive tool catalog, configuration, safety model, errors |
| Explanation | Architecture, auth model, safety rationale, rate limiting |
| Development | Architecture, contributing, testing |
Development
pnpm typecheck # TypeScript strict checks
pnpm lint # oxlint
pnpm test # vitest (194 tests, 87% coverage)
pnpm build # tsup → dist/
License
MIT © The Self Agency LLC
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。