Agent-orchestrating multiplexer

The inversion of the agent-native-multiplexer: not the multiplexer managing agents, but an agent driving the multiplexer — via herder's official agent-skill-file, an agent can create workspaces/sessions, start other agents inside them, prompt them, and read their replies. alejandro-ao's demo: Tau receives "create a new Herder session, run Pi inside of it and ask it to say hello world" and executes the full chain. Effectively a DIY sub-agent runtime built from a terminal tool — the hobbyist edge of what dhh does with his bot system and what harnesses do natively.

Source: report