Agent control plane
The layer that starts, watches, limits and stops many AI agents at once: queues, permissions, budgets, logs and a view of who is doing what.
One engineer can supervise a few agent sessions in terminals. Ten parallel background agents across several repositories need what infrastructure teams call a control plane: task assignment, per-agent permissions and secrets, spending limits, status and traces, and a way to pause or kill a run. Coding tool vendors, cloud providers and agent platforms are building it in different forms. Without it, a software factory has no foreman.
Related terms
- Multi-agent system: Several AI agents with separate roles and contexts working on one task, usually coordinated by a lead agent or a fixed workflow.
- Background coding agent: A coding agent that runs in the cloud on its own, takes an issue or a task and comes back later with a pull request for review.
- Cost per task (agent unit economics): What one finished agent task really costs: tokens, retries, tool calls and the human time to review it, compared with doing it another way.
- Agent harness (harness engineering): Everything around the model that makes it an agent: the loop, tools, permissions, memory, context management.
- Software factory (AI agents): A pipeline where agents write and fix most of the code while people write specs, set guardrails, design the checks and decide what ships.
