Systems Architect
Takes a plain-language prompt and produces a structured plan with agent assignments, skill requirements, and integration mappings.
Ren AIConnects to
Systems Architect turns a vague idea into an executable plan. You describe what you want to automate in plain English. The agent researches the problem, checks the registry for existing agents and skills that fit, and designs a multi-agent system that covers the workflow end to end.
The output is a structured execution plan: which agents to use, what skills each needs, which MCP integrations to connect, and the step-by-step execution order. The plan is formatted so the Build agent can pick it up and deploy everything directly. You review and edit the plan before anything gets built. If your prompt is ambiguous, the agent asks clarifying questions before committing to an architecture.
How it works
Parse the natural-language prompt
Reads your description of what you want automated. Identifies the core goals, the services involved, and any constraints you mentioned. Asks clarifying questions if the prompt is ambiguous.
Search the registry for existing components
Checks the Ren registry for agents, skills, and workflows that already cover parts of the requested automation. Reuses existing components to avoid duplication.
Design the multi-agent system
Maps out which agents handle which part of the workflow, what skills each agent needs, and which MCP integrations must be connected. Resolves dependency order between agents.
Output the execution plan
Produces a structured plan with agent assignments, skill specs, integration mappings, and step-by-step execution order. The plan is ready for review and can be passed directly to the Build agent.
FAQs
A structured plan listing each agent, its skills, its MCP integrations, and the execution order. The format is directly consumable by the Build agent.
Yes. The architect produces a plan and waits for your approval. You can modify agent assignments, swap skills, or adjust the execution order before handing it to Build.
It checks the registry first. If an existing agent or skill fits the task, it includes that in the plan instead of proposing a new one.
It asks clarifying questions. It will not commit to an architecture until it has enough information to produce a plan that actually covers your workflow.