Agent runtime
One runtime runs every agent. It supplies the model and the tool-calling loop; DomeCommand supplies the tools, the surfaces a draft lands on, and the credential.
The runtime is behind DOME AI, the Copilot, the scenario composer and the rule drafter. What each of them does, and what none of them can do, is on Agentic C2.
If the runtime is unavailable
The chat panel goes quiet and everything else works. The air picture, fusion, the planner, the ladder and the line are in the server, and none of them ask an agent anything. An air picture that stopped updating because a language model timed out would be worse than no assistant.
There is no fallback. A surface that uses a model says it is unavailable and contributes nothing. It does not substitute a canned answer. Off, broken and unconfigured are three states, and the console tells them apart.
How the console reaches it
The console never holds an API key. It asks the DomeCommand server for a short-lived token, which the server mints from its own credential and refreshes on a cadence. The key stays server-side, a browser session carries something that expires, and access is revoked once in the server config rather than per user.
Configuration
A deployment needs a runtime workspace, an agent in it, and a credential. Set those on the server and the chat panel comes up.
Where to go next
- Agentic C2 for what you can ask it and what it drafts.
- Policy & rules for the drafting flow end to end.