What Are Autonomous Agents?

#073 0:18 Part of the Rysh video series

Headless AI workers. No terminal, no pane -- just a brain with tools.

What Are Autonomous Agents? — 0:18 walkthrough

What you'll see

Launch rysh; explain every pane is shell + agent, but autonomous agents are headless.
Double-Escape into rysh mode (## prompt).
##agent spawn code-reviewer "You review code for quality, correctness, and best practices." -- no PTY, its own AgenticActor.
##agent list shows each agent and its status.
@code-reviewer summarize what a good code review checks for -- prompt it with the @name prefix; it runs with full tool access in the background.

Commands shown

rysh
##agent spawn code-reviewer 'You review code for quality, correctness, and best practices.'
##agent list
@code-reviewer summarize what a good code review checks for

Keys used

EnterEscape

Transcript

0:00Headless AI workers. No terminal, no pane -- just a brain with tools.

0:03Launch rysh. Every pane is a shell plus an agent -- but autonomous agents live entirely in the background.

0:08Double-press Escape twice to reach rysh mode, where the double-hash system commands live.

0:14Spawn an agent with a name and a system prompt. It has no PTY and no terminal -- just its own AgenticActor running independently in the workspace.

0:24Check ##agent list to see every agent and its status. Created via ##agent spawn, prompted with @name, controlled with @@name.

0:32Talk to it with the at-sign prefix from any input mode. The agent runs the prompt with full tool access in the background.

0:44Autonomous agents: tireless workers you spawn, prompt, and command -- without ever opening a pane.

Key takeaway

Autonomous agents are headless actors with their own AgenticActor -- created with ##agent spawn, prompted with @name, controlled with @@name.