v0.61.0

Added

• dl --ls says which workspaces run an older image than their reference names (#673).
  A pull moves an image reference and leaves every running container on the image it was
  created from, so long-lived workspaces drifted days behind and nothing said so. dl --ls
  now prints one line under the table for each such workspace, naming the reference and
  dl <ws> recreate, and dl --ls --json appends a staleImage key to its row. The check
  reads devpod's workspace_result.json and asks docker for the container's image and the
  reference's current one; where devpod built a derived image on top of the declared one,
  it compares layers. Every doubt reads as current.
• dl <ws> recreate starts the agents it ends again, in the panes they held (#673).
  A recreate ended every agent session in the container, which is why a stale workspace
  was left alone. In a herdr pane, the recreate now reads which panes hold a live agent in
  the workspace and the line herdr saved to start each one again
  (dl <ws> -- ... claude ... --resume <uuid>). Once the new container is up, it types
  each line back into its pane with herdr pane run, and then attaches. For a pane that
  will not take its line, it prints the line to run by hand. Outside herdr there is no
  line to print, so it says that the recreate ends the agents and that claude --resume
  in the new container picks a conversation up again. No failure here refuses the
  recreate. See docs/cli.md, "A recreate starts the agents it ends again".
• dl --refresh-stale recreates the stale workspaces that are idle (#673). It runs
  the recreate above, with no attach, on each workspace dl --ls calls stale, so it can
  run after the daily pull. It skips a workspace, and says why, when another dl is
  launching it, a build runs in it, an agent (claude, codex or gemini) runs a
  command in it or is not idle, or it holds an agent session dl cannot start again
  (outside herdr, any agent session). It prints its plan first and a count last, checks
  each workspace again just before its recreate, and exits 1 when a recreate failed.
  See docs/cli.md, "--refresh-stale: recreate the stale workspaces that are
  idle".