Dobby MCP
- dobby-ai/dobby
Dobby is an MCP server that lets coding agents connect to the Dobby platform to manage tasks, approvals, schedules, and governance for AI workers.
About
Dobby is a Model Context Protocol server that connects Claude Code, Cursor, and similar agents to the Dobby platform for operational control of AI workers. developers who run multiple automations or customer-facing agents use it when they need a single place to assign tasks, watch execution, pause risky runs, and route sensitive steps through approvals without building custom admin UIs. The server exposes platform capabilities over stdio or hosted HTTP so your agent can list work, trigger schedules, and respect tenant governance using API keys tied to a workspace. It fits teams who already centralize agents on Dobby and want MCP-native wiring instead of copying dashboard clicks into prompts. It does not replace your agent’s coding skills; it adds a control plane. Version 1.1.0 is declared in the MCP registry manifest with repository at github.com/dobby-ai/dobby.
- stdio MCP via npm @dobbyai/mcp-external (v1.1.0) or remote streamable-http at https://mcp.dobby-ai.com/mcp
- Task and schedule management for AI agents from the agent session
- Human-in-the-loop approvals and governance hooks for agent actions
- Requires DOBBY_API_KEY (dk_* or gk_*) and DOBBY_TENANT_ID; optional DOBBY_API_URL
- MCP integration to the Dobby platform—not a standalone agent skill or local script
Dobby MCP by the numbers
- Data as of Aug 10, 2026 (Skillselion catalog sync)
claude mcp add --env DOBBY_API_URL=YOUR_DOBBY_API_URL --env DOBBY_API_KEY=YOUR_DOBBY_API_KEY --env DOBBY_TENANT_ID=YOUR_DOBBY_TENANT_ID dobby -- npx -y @dobbyai/mcp-externalAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| Package | @dobbyai/mcp-external |
|---|---|
| Transport | STDIO, HTTP |
| Auth | Required |
| Repository | dobby-ai/dobby ↗ |
What it does
Wire your coding agent to Dobby so you can queue tasks, require human approvals, run schedules, and enforce governance from the IDE instead of a separate ops console.
Who is it for?
Best when you're already on Dobby and want MCP-driven task queues, schedules, and approval flows from Claude Code or Cursor.
Skip if: Skip if you only need local file edits and have no Dobby workspace, or anyone and wants payments or AWS emulation instead of agent ops.
What you get
After you register the MCP server with your API key and tenant ID, your agent can orchestrate and monitor Dobby-managed work from the same session where you code and debug.
- Live MCP connection to your Dobby tenant for agent task and schedule control
- Approval and governance actions callable from the agent session
- Optional remote MCP endpoint without local npm install
By the numbers
- Registry version 1.1.0
- Two transports: npm stdio package and remote streamable-http URL
- Three documented environment variables (two required)
Recommended MCP Servers
Agent Audit MCPImmutable audit trail for agent-to-agent interactions2
Agentguard47 MCPRead-only MCP server for coding-agent traces, alerts, costs, usage, and budget health.4
Agentic Observability MCPAgent tracing, cost tracking, anomaly detection for LLM agents
Agent Incident Relay MCPAgent Incident Relay MCP — Article 73 5-clock broadcaster.
Agent Infra MCP18 tools: agent memory, API drift detection, and HMAC-signed trust receipts
Agentmonitorrelay MCPAI agent run monitoring with incident replay and SLA receipts.How it compares
Hosted agent control-plane MCP, not a local dev tool or payment rail.
FAQ
Who is Dobby MCP MCP for?
Developers and small teams using the Dobby MCP platform who want their coding agent to assign tasks, monitor runs, and enforce approvals without leaving the IDE.
When should I use Dobby MCP MCP?
Use it in the Operate phase when agents run recurring or production workflows and you need governance, scheduling, and visibility from MCP-capable clients.
How do I add Dobby MCP MCP to my agent?
Install @dobbyai/mcp-external from npm for stdio, set DOBBY_API_KEY and DOBBY_TENANT_ID (and DOBBY_API_URL if self-hosted), or point your client at the remote URL https://mcp.dobby-ai.com/mcp for streamable HTTP.