Architect -- AI Agent System Designer

Most AI agents fail in production because nobody designed the system they run in.

You can build a brilliant single agent, but the moment you need agents talking to each other, sharing memory, or handing off tasks -- things collapse. Skill conflicts multiply. Context windows overflow. Orchestration turns into spaghetti. And debugging a multi-agent system without architecture? Good luck.

Architect is the system designer your agent fleet needs. It does not write code -- it designs the systems that make code work. Think of it as the principal engineer who draws the blueprints before anyone touches a keyboard.

What You Get:

• Agent System Design -- map complete architectures with capability matrices, communication protocols, and failure boundaries
• Skill Graph Mapping -- visualize skill dependencies, detect conflicts, and plan capability coverage across your entire agent fleet
• Orchestration Patterns -- design coordination strategies for multi-agent workflows with proper handoff, delegation, and escalation
• Memory Architecture -- plan context management, shared state, and persistent knowledge systems that scale
• Agent Testing Strategy -- build comprehensive test plans covering capability verification, integration testing, and chaos scenarios

Pairs well with Overlord for hands-on agent operations after the architecture is set.

Works with OpenClaw and any agent setup. Drop in and go.