# Dynamic Graph

> Agents can create, clone, activate or put to sleep other Agents while the work runs, growing their group as they go.

Inside a Graph, an Agent whose role holds the graph tools can create, clone, activate or put to sleep other Agents. This lets a group grow while it works instead of keeping the shape it started with.

Concretely, an Agent can create a peer from an existing role, clone a running Agent (its configuration, context and tools) into a new identity, copy the whole graph with `clone_graph` (every sub-Graph and every Agent's context included), or run another template as a sub-Graph and act as its own human, answering the sub-Graph's questions itself. In The One today the main assistant does not grow your work this way: it gives each matter's work to a working group, a run of its own. The growing happens inside the groups, where a team lead clones workers, starts TaskBoards or runs sub-Graphs.

This capability set, an Agent growing the Graph at runtime, is what we call [Prolifera Engineering](../prolifera-engineering). A few invariants are worth stating plainly. A cloned or created Agent is **born disconnected and inactive**: it must be wired in and activated before it does anything, and a copied graph starts dormant. An Agent creates or clones only when it **chooses** to call the tool; a long context does not trigger it, which is handled separately by [Meditation](./meditation). And an Agent can do this at all only if its role holds the founding tools (`create_subgraph`, `clone_agent_node`, `create_agent_node`, `clone_graph`, `create_task_board`). Because the decision is the Agent's own and nothing caps it automatically, such an Agent can fan out without limit; the authority is concentrated and observable by design, not prevented. A usage limit you set on its matter or run pauses it, and **Stop All Agents**, in the app menu or the tray menu, stops everything.

Source: https://the-zeroth.com/en/docs/multi-agent-architecture/more-possibilities/dynamic-graph