A development kit for building custom plugins that extend OpenClaw agent capabilities with new tools, integrations, and behaviors.
The Plugin SDK lets you extend OpenClaw beyond its built-in capabilities. Build custom MCP servers, add new tool integrations, create specialized agent behaviors, and package them as reusable plugins.
Plugins follow the MCP standard, so anything you build works with any MCP-compatible agent — not just OpenClaw. The SDK provides helpers for authentication, error handling, input validation, and testing.
Common plugin use cases: connecting internal APIs, adding industry-specific tools, building custom approval workflows, and integrating with proprietary systems that have no public MCP server.
Every organization has unique tools and workflows. The Plugin SDK means you are not limited to pre-built integrations. Extend your agent to work with any system in your stack.
Clawctl deploys custom plugins alongside your agent. Plugins run in the same sandbox with the same security controls. The dashboard shows plugin health and tool call metrics.
Try Clawctl — 60 Second DeployTypeScript and Python are the primary supported languages. Any language that can serve HTTP/stdio can work with MCP.
Yes. Plugins run inside the agent sandbox with the same egress filtering, audit logging, and approval workflows.
Yes. Plugins follow the open MCP standard and can be shared publicly or within your organization.
MCP Server
A service that exposes tools and data to an AI agent through the Model Context Protocol, enabling standardized integrations with external systems.
MCP (Model Context Protocol)
An open protocol that lets AI agents connect to external tools and data sources through a standardized interface.
Tool Use
The ability of an AI agent to interact with external tools and APIs — reading data, calling functions, and taking actions in the real world.
OpenClaw Architecture
The technical design of OpenClaw: a gateway that routes messages from 23+ channels to an LLM-powered agent with MCP tool integrations, running in a Docker sandbox.