Agent Orchestrator MCP Server

> Note: This package is part of the MCP Servers monorepo. For the latest updates and full source code, visit the Agent Orchestrator MCP Server directory.

MCP server for PulseMCP's agent-orchestrator: a Claude Code + MCP-powered agent-parallelization system for agentic coding and ops done at PulseMCP.

Highlights

- Simplified 5-tool interface for full agent session management
- Search, filter, and retrieve sessions with optional logs and transcripts
- Session lifecycle actions (pause, restart, archive, unarchive, follow_up, change_mcp_servers)
- Static configuration access via tools and MCP resources
- Tool grouping system for permission-based access control
- TypeScript with strict type checking
- Comprehensive testing setup (functional, integration, manual)

Capabilities

$3

| Tool | Group | Description |
| ----------------- | -------- | ---------------------------------------------------------------------------------- |
| search_sessions | readonly | Search/list sessions with optional ID lookup, query, and status filter |
| get_session | readonly | Get detailed session info with optional logs and transcripts |
| get_configs | readonly | Fetch all static configuration (MCP servers, agent roots, stop conditions) |
| start_session | write | Create and start a new agent session |
| action_session | write | Perform actions: follow_up, pause, restart, archive, unarchive, change_mcp_servers |

$3

| Resource | Description |
| ---------------------------------------------- | ----------------------------------------------- |
| agent-orchestrator://config | Server configuration and status (for debugging) |
| agent-orchestrator://configs/mcp-servers | List of available MCP servers for sessions |
| agent-orchestrator://configs/agent-roots | Preconfigured repository settings with defaults |
| agent-orchestrator://configs/stop-conditions | Session completion criteria definitions |

$3

Control which tools are available via the ENABLED_TOOLGROUPS environment variable:

| Group | Description |
| ---------- | --------------------------------------------------------------------------- |
| readonly | Read-only operations (list, get, search) |
| write | Write operations (create, update, follow_up, pause, restart, archive, etc.) |
| admin | Administrative operations (delete) |

Examples:

- ENABLED_TOOLGROUPS="readonly" - Only read operations
- ENABLED_TOOLGROUPS="readonly,write" - Read and write, no admin
- Not set - All tools enabled (default)

Setup

$3

- Node.js (use nvm use if you have nvm installed)
- An Agent Orchestrator instance with API access

$3

| Variable | Required | Description | Default |
| ----------------------------- | -------- | -------------------------------------- | ----------- |
| AGENT_ORCHESTRATOR_BASE_URL | Yes | Base URL for the orchestrator API | - |
| AGENT_ORCHESTRATOR_API_KEY | Yes | API key for authentication | - |
| ENABLED_TOOLGROUPS | No | Comma-separated tool groups | All enabled |
| SKIP_HEALTH_CHECKS | No | Skip API connectivity check at startup | false |
| HEALTH_CHECK_TIMEOUT | No | Health check timeout in milliseconds | 10000 |

$3

Make sure you have your Agent Orchestrator base URL and API key ready.

Then proceed to the setup instructions below. If this is your first time using MCP Servers, you'll want to make sure you have the Claude Desktop application and follow the official MCP setup instructions.

#### Manual Setup

You're going to need Node working on your machine so you can run npx commands in your terminal. If you don't have Node, you can install it from nodejs.org.

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

Modify your claude_desktop_config.json file to add the following:

``json
{
"mcpServers": {
"agent-orchestrator": {
"command": "npx",
"args": ["-y", "agent-orchestrator-mcp-server"],
"env": {
"AGENT_ORCHESTRATOR_BASE_URL": "http://localhost:3000",
"AGENT_ORCHESTRATOR_API_KEY": "your-api-key-here",
"ENABLED_TOOLGROUPS": "readonly,write"
}
}
}
}
`

Restart Claude Desktop and you should be ready to go!

Development

$3

`bash
npm run dev
`

$3

`bash


Run functional tests in watch mode

npm test

Run tests once

npm run test:run

Run integration tests (full MCP protocol)

npm run test:integration

Run all automated tests

npm run test:all
`

$3

`bash


Check for linting issues

npm run lint

Auto-fix linting issues

npm run lint:fix

Format code

npm run format
`

API Coverage

This MCP server provides tools for the following Agent Orchestrator REST API endpoints:

$3

- GET /api/v1/sessions - List sessions
- GET /api/v1/sessions/search - Search sessions
- GET /api/v1/sessions/:id - Get session
- POST /api/v1/sessions - Create session
- PATCH /api/v1/sessions/:id - Update session
- DELETE /api/v1/sessions/:id - Delete session
- POST /api/v1/sessions/:id/archive - Archive session
- POST /api/v1/sessions/:id/unarchive - Unarchive session
- POST /api/v1/sessions/:id/follow_up - Send follow-up prompt
- POST /api/v1/sessions/:id/pause - Pause session
- POST /api/v1/sessions/:id/restart - Restart session

$3

- GET /api/v1/sessions/:session_id/logs - List logs
- POST /api/v1/sessions/:session_id/logs - Create log

$3

- GET /api/v1/sessions/:session_id/subagent_transcripts - List transcripts
- GET /api/v1/sessions/:session_id/subagent_transcripts/:id` - Get transcript

License

MIT