@tsdevstack/cli-mcp
MCP server plugin for the tsdevstack CLI, enabling AI agents to manage infrastructure, deployment, and project state with 54 tools and 12 resources.
README
@tsdevstack/cli-mcp
MCP (Model Context Protocol) server plugin for the tsdevstack CLI. Exposes 54 tools and 12 resources so AI agents like Claude can understand and assist you with the framework, the infrastructure, deployment flows and querying the project state.
Features
- 54 MCP Tools — 13 read-only queries + 41 actions covering the full tsdevstack CLI surface
- 12 MCP Resources — project state, secrets context, Kong routes, and framework guides
- Stdio Transport — connects via stdin/stdout for IDE and agent integration
- Zod Validation — all tool inputs validated with schemas
- MCP Annotations — tools include
readOnlyHint,destructiveHint, andidempotentHintmetadata
Installation
npm install @tsdevstack/cli-mcp
Requires tsdevstack as a peer dependency.
Usage
As a CLI plugin
The MCP server registers as a plugin with the tsdevstack CLI:
import { initContext, registerMcpPlugin } from '@tsdevstack/cli-mcp';
// Initialize with CLI plugin context
initContext(pluginContext);
// Register the mcp:serve command
registerMcpPlugin(program);
Starting the server
npx tsdevstack mcp:serve
This starts the MCP server over stdio transport, ready for AI agent connections.
Claude Code integration
Add to your .mcp.json:
{
"mcpServers": {
"tsdevstack": {
"command": "npx",
"args": ["tsdevstack", "mcp:serve"]
}
}
}
Tools
Query tools (read-only)
| Tool | Description |
|---|---|
list_services |
List all services with types and ports |
list_environments |
List configured cloud environments |
get_project_config |
Full project configuration |
get_infrastructure_config |
Per-environment infrastructure settings |
get_service_status |
Cloud resource status for a service |
list_deployed_services |
All deployed services in an environment |
list_secrets |
Secret names in cloud (not values) |
diff_secrets |
Compare local vs cloud secrets |
get_secret |
Get a single secret value |
list_schedulers |
Scheduled jobs and status |
plan_db_migrate |
Preview pending database migrations |
infra_plan |
Terraform plan preview |
infra_status |
Infrastructure sync status |
Action tools — local
| Tool | Description |
|---|---|
sync |
Regenerate all local config |
generate_secrets |
Regenerate local secrets files |
generate_kong |
Regenerate Kong gateway config |
generate_docker_compose |
Regenerate docker-compose.yml |
add_service |
Add a new service (nestjs, nextjs, spa) |
remove_service |
Remove a service from the project |
generate_client |
Generate TypeScript HTTP client from OpenAPI |
add_bucket_storage |
Add an object storage bucket |
remove_bucket_storage |
Remove an object storage bucket |
add_messaging_topic |
Add a messaging topic to config |
remove_messaging_topic |
Remove a messaging topic from config |
update_messaging_topic |
Update publishers/subscribers for a topic |
register_detached_worker |
Register a detached worker in config |
unregister_detached_worker |
Remove a worker from config |
Action tools — cloud
| Tool | Description |
|---|---|
cloud_secrets_push |
Push secrets to cloud |
cloud_secrets_set |
Set a single cloud secret |
cloud_secrets_remove |
Remove a cloud secret |
infra_deploy |
Full infrastructure deployment |
deploy_services |
Deploy code changes to services |
deploy_service |
Build + push + deploy a single service |
deploy_kong |
Rebuild and deploy Kong gateway |
deploy_lb |
Deploy/update load balancer |
run_db_migrate |
Apply pending database migrations |
deploy_schedulers |
Deploy all scheduled jobs |
deploy_scheduler |
Deploy a single scheduled job |
remove_service_cloud |
Remove a service from cloud |
remove_detached_worker |
Remove a worker from cloud |
infra_destroy |
Destroy all cloud infrastructure |
Action tools — setup & CI
| Tool | Description |
|---|---|
cloud_init |
Initialize cloud provider credentials |
infra_bootstrap |
Bootstrap cloud project |
infra_init |
Initialize Terraform state |
infra_generate |
Generate Terraform files |
infra_generate_docker |
Generate Dockerfiles |
infra_build_docker |
Build Docker images |
infra_push_docker |
Push Docker images to registry |
infra_generate_kong |
Generate Kong declarative config |
infra_build_kong |
Build Kong Docker image |
infra_init_ci |
Initialize CI/CD workflows |
infra_generate_ci |
Regenerate CI workflows |
validate_service |
Validate service structure |
remove_scheduler |
Remove a scheduled job from cloud |
Resources
| Resource | Description |
|---|---|
| Project state (8) | Services list, environments, config, infrastructure settings, deployed services, secrets, schedulers, migration status |
| Secrets context | Local and cloud secret names with diff |
| Kong routes | Current gateway routing configuration |
| Guides (4) | Framework guide, workflow guide, nest-common guide, config guide |
Architecture
The MCP server is a thin layer over the tsdevstack CLI:
- Query tools either read local config files directly or wrap
npx tsdevstackcommands - Action tools delegate to
runCommand()which spawnsnpx tsdevstack <command>as a child process - Resources read project files and provide contextual documentation
Community
Join the Discord: discord.gg/2EMFkqc8QR
License
Apache-2.0
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。