nyldn/claude-octopus v11.15.4

on GitHub

Fixed

• A probe seat now runs for the OCTOPUS_AGENT_TIMEOUT budget. orchestrate.sh probe gave each seat the 600-second TIMEOUT default even with
  OCTOPUS_AGENT_TIMEOUT=1200 set, and the timeout message then recommended
  setting the variable that was already set. Probe seats, probe-single and
  the probe progress watchdog now use the budget run_agent_sync uses:
  OCTOPUS_AGENT_TIMEOUT, then --timeout, then the default. A timed-out
  probe seat recommends OCTOPUS_AGENT_TIMEOUT when it is set. Other
  background workflows still wait for their seats against --timeout, so
  their seats keep that budget.
  #1241,
  commit 1f8d395.

Full changelog: v11.15.3...v11.15.4