mawkone
b6d7148ded
- Control Plane API with Gemini integration - Executors: Deploy, Analytics, Marketing - MCP Adapter for Continue integration - VSCode/VSCodium extension - Tool registry and run tracking - In-memory storage for local dev - Terraform infrastructure setup
Product OS MCP Adapter
Exposes Control Plane tools to Continue and other MCP clients.
Architecture
Continue (MCP client)
↓
MCP Adapter (this server)
↓
Control Plane API
↓
Gemini + Executors
Available Tools
| Tool | Description |
|---|---|
deploy_service |
Deploy a Cloud Run service |
get_service_status |
Check deployment health |
get_funnel_analytics |
Get conversion metrics |
get_top_drivers |
Understand metric drivers |
generate_marketing_posts |
Create social content |
chat_with_gemini |
General AI conversation |
Setup
1. Install dependencies
cd platform/backend/mcp-adapter
npm install
2. Make sure Control Plane is running
cd platform/backend/control-plane
npm run dev
3. Install Continue Extension
In VS Code/VSCodium:
- Open Extensions (Cmd+Shift+X)
- Search for "Continue"
- Install
4. Configure Continue
The configuration is already in .continue/config.yaml.
Or manually add to your Continue config:
experimental:
modelContextProtocolServers:
- name: productos
command: npx
args:
- tsx
- /path/to/platform/backend/mcp-adapter/src/index.ts
env:
CONTROL_PLANE_URL: http://localhost:8080
5. Use in Continue
- Open Continue chat (Cmd+L)
- Enable Agent mode (click the robot icon)
- Ask: "Deploy my service to staging"
Environment Variables
| Variable | Default | Description |
|---|---|---|
CONTROL_PLANE_URL |
http://localhost:8080 |
Control Plane API URL |
TENANT_ID |
t_mcp |
Tenant ID for tool calls |
Development
# Run directly (for testing)
npm run dev
# Build
npm run build
# Run built version
npm start
Testing
# Test tool listing
echo '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' | npm run dev
# Test tool call
echo '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"get_funnel_analytics","arguments":{}}}' | npm run dev