Run the CLI

The core-agent binary supports two interaction modes. Pick the one that matches what you’re trying to do.

Getting startedInstall, point at a provider, first turn — under five minutes.

Interactive (TUI)You're driving the agent. Bubble Tea TUI, slash commands, conversation history preserved. For code review, exploration, ad-hoc tasks.

Autonomous (headless)The agent runs unattended against a goal — monitors, scheduled jobs, remediation workers. Budgets cap turns/tokens/cost/wallclock.

Skills libraryThree bundled Claude-Skills that walk you through CLI setup, autonomous setup, and library embedding.