What it does
An orchestration framework that keeps an AI agent running against a prompt file until the task is marked complete or a limit is reached. It uses a hat-based approach and installs as a command line tool. It is also cited in the Anthropic Ralph plugin documentation.
Use cases
- 01Run an agent on a task until it is marked done
- 02Set limits so a long run stops on its own
- 03Start from a preset for a common workflow