Agents

Manage your AI workforce — Executive, Manager, and Worker agents.

Agents are the core actors. Each has a level, mission, model settings, and budget policy. The Executive Agent (L0) manages Manager Agents (L1) which manage Workers (L2).

Views

• Agent list — all agents with status and level.
• Agent detail — mission, manager, skills, recent runs, budget usage.
• Hierarchy tree — visualize manager/worker relationships.
• Create / edit form — name, title, level, mission, model, budget policy.

Actions

• Create an agent
• Edit
• Archive
• Pause / resume
• Assign manager
• Assign skills
• Inspect budget usage

Status lifecycle

active · idle · running · error · paused · terminated · archived

Terminated or archived agents cannot receive new tasks. Paused agents are skipped by the heartbeat scheduler.

Levels and routing

Level	Role
L0	Executive Agent — top of hierarchy. Exactly one active per company.
L1	Manager Agent — owns a domain or department.
L2	Worker Agent — executes specialized tasks.

Route access in the API is gated by minimum level; tool lists are filtered per session by the recipient agent’s level.

Sub-agent creation

The Executive Agent can propose or create subordinate agents. Creation follows the approval policy — by default, agent creation requires Founder approval unless explicit policy says otherwise.