First Agent

Agents in matic are persistent, stateful actors rather than stateless tool invocations. Each agent is spawned from an Archetype, lives in its own filesystem directory within the org repo, and evolves through a defined lifecycle that includes onboarding, work, learning, and assessment. This section walks through what it takes to stand up your first agent, from choosing the right blueprint to understanding the operating systems that make the agent usable in practice.

Choose an Archetype

How to select the right Archetype blueprint for your agent, the reusable template that defines its persona, skills, and initial capability profile. Read more

Spawn Agent

The spawning flow from request through Auto Matic validation, directory scaffolding, and asset loading, covering everything that happens before onboarding begins. Read more

Agent Onboarding Sequence

The five-step onboarding process every agent completes before becoming available: Charter Read, Library Ingest, Probe Initialisation, Charter Alignment Check, and Memory Initialisation. Read more

Agent Lifecycle States

The full state machine governing an agent's operational life, from Spawned through Available, Engaged, Learning, Assessment, Idle, Suspended, and Decommissioned. Read more

Agent Workspace and Memory

How an agent's filesystem directory is structured and how the four knowledge stores, Memory, Experience, Library, and Decisions, serve different purposes during execution. Read more

Probes and Skills

Probes are lightweight heuristics that describe what an agent can do, how confident it is, and what it will not do. Skills are structured capability specifications that give tools operational context; together they define the agent's competence profile. Read more

Auto Matic - the Chief Agent

Every org has exactly one Auto Matic instance, the chief agent that routes signals, validates spawn requests against the Charter, and coordinates org-level planning. Read more