Core Concepts
An overview of the foundational concepts in WorkClaw -- Claws, Workspaces, Skills, Connections, and Teams. Understand how the pieces fit together.
What are the core concepts in WorkClaw?
WorkClaw is built around five foundational concepts. Each one plays a specific role in how your AI agents operate, what they can do, and who controls them.
What is a Claw?
A Claw is an AI agent -- your digital teammate. Every team starts with a System Claw that is shared and ready to use. You can also create custom Claws for specific roles, workflows, or team members. Each Claw has its own identity, skills, connections, and workspace.
What is a Workspace?
A Workspace is the isolated environment where a Claw operates. It contains the Claw's configuration files (AGENTS.md, SOUL.md, MEMORY.md, USER.md), installed skills, uploaded files, and session history. Workspaces ensure that each Claw has its own context and memory, separate from other agents on the team.
What is a Skill?
A Skill is a modular capability package that expands what a Claw can do. Skills come in three types: system (built-in and maintained by WorkClaw), third-party (from the marketplace), and team (custom skills created by your organization). You install skills per-Claw, so each agent can have a different set of capabilities.
What is a Connection?
A Connection is a secure integration with an external service. Connections let your Claw interact with tools like Google Workspace, Slack, and custom APIs. They are authenticated via OAuth or API credentials and can be scoped as private (personal) or team-wide (shared).
What is a Team?
A Team is the organizational container for everything in WorkClaw. It holds your Claws, members, connections, and billing. Teams have role-based access control with three roles: Owner, Admin, and Member. The team determines who can configure agents, invite members, and manage integrations.
How do these concepts fit together?
The relationship between concepts follows a clear hierarchy:
- A Team contains members, Claws, and shared connections.
- Each Claw belongs to a team and has its own Workspace.
- Skills are installed per-Claw and define what the agent can do.
- Connections link a Claw to external services and are registered per-Claw but can be shared at the team level.
This architecture means you can have multiple specialized agents on one team, each with different capabilities and integrations, all managed by the same group of people with appropriate role-based permissions.