The fleet screen

Run tiny with no arguments:

    NAME        STATE      AGE   CPU    MEM  WHAT
  ✳ s-x7k2f     needs you  12m   84m  512Mi  May I force-push the rebased branch?
  ● api-fix     running     3h  212m  1.1Gi  migrating auth tests to vitest
  └ ● api-db    running    41m  907m  2.9Gi  rewriting migrations in golang:1.26
  ● night-run   running     8h    1m  301Mi  ⏸ Usage limit reached · continuing at 5:20pm
  ✓ readme      done        2d
  · ✉ broadcast to all…
  · ＋ new session
  · ⚙ new session with options…
  · ☰ namespace settings
  · ✕ quit

Reading it

WHAT is live. It shows the agent's own declared title, refreshed as it works. A paused session says why and when it resumes; a session with a bad token quotes the agent's own error message.
CPU/MEM are self-reported by each session from its own cgroup, so the screen works without metrics-server or extra RBAC.
✳ means a question waits. Rows keep creation order on purpose, so nothing jumps around when a question arrives.
Children render under their parent (└).

Keys

key	what
`enter`	attach (also `tiny attach <session>` from a plain shell); on an idle "waiting for you" row, attaching clears the mark
`a`	answer the selected ✳ — answering is acting
`m`	message the selected session (durable inbox)
`b`	broadcast to every unfinished session
`n` / `o`	new session — blank / the full options form
`d`, then `y`	delete (two keystrokes on purpose: a session is a workspace and a transcript)
`r`	refresh

Drop a file onto the terminal — fleet screen or attached — and it streams to /workspace/uploads/ with live progress; the agent gets the path.

Namespace settings

☰ toggles the add-ons: registry cache, artifact store, GitHub runner. Your client applies each one with your credentials when you flip it.