Agents

Clawpy organises its AI workers into specialised archetypes — each with a defined role, skillset, and workspace. Instead of one general-purpose assistant, you get a team of experts that collaborate.

Agent Archetypes

The Architect

Designs system specifications and handles high-level structural planning. The Architect analyses requirements, proposes solutions, and creates technical blueprints before any code is written.

Best for: System design, architecture decisions, project planning, specification writing.

The Builder

Translates specifications into clean, functional code. The Builder works from the Architect's plans and focuses on implementation quality, testing, and documentation.

Best for: Code generation, feature implementation, test writing, refactoring.

The Implementor

Executes deep operational tasks, deployment scripts, and terminal commands. The Implementor handles the hands-on work of shipping — CI/CD pipelines, server configuration, and infrastructure management.

Best for: DevOps, deployment, shell scripting, infrastructure automation.

The Auditor

Reviews work produced by other agents. The Auditor validates code quality, checks for security issues, and ensures outputs meet the original specification. The Auditor holds veto power over deployments.

Best for: Code review, security auditing, quality assurance, compliance checking.

How Agents Collaborate

When you send a task through the operator chat, the swarm conductor analyses your request and decides which agents to involve:

1. Complex tasks may start with the Architect for planning, then move to the Builder for implementation, and finish with the Auditor for review.
2. Simple tasks may go directly to a single agent.
3. Agents can delegate sub-tasks to each other when they encounter work outside their specialisation.

All collaboration happens through the Agent Communication Protocol (ACP), which provides a traceable message bus between agents. See Agent Communication for details.

Workspaces

Every agent operates within its own isolated workspace — a dedicated directory containing its files, memory, and configuration. Agents cannot access each other's workspaces unless explicitly granted collaboration access.

This isolation ensures that a bug in one agent's work never corrupts another agent's state.

Custom Agents

You can create custom agent archetypes by defining a soul.md file that specifies:

• The agent's role and personality
• Which tools it can access
• Its memory retention policy
• Budget and rate limits