[Bug] Resume session returns to Fleetview dashboard and freezes application

Resolved 💬 1 comment Opened Jul 8, 2026 by gael64 Closed Jul 8, 2026

Bug Description
Since this morning I am not able to resume a session as the resume option always returns the Fleetview dashboard and no action is possible from there, there dahboard gets frozen, the only way to get out is to close the terminal.

Environment Info

  • Platform: darwin
  • Terminal: Apple_Terminal
  • Version: 2.1.204
  • Feedback ID: 3cbeaa7f-b9cc-466a-b02c-e7cd55be9b3a

Claude Code v2.1.204 (bug present since at least v2.1.202)

Job: 779aeb30-ddab-4071-9383-216788254e7b ("Weather station pilot project setup", bg template, computer-use tools + claude-fable-5[1m])

  • state.json said "state": "done" from 2026-07-07T17:21:33Z onward
  • but pids 40061 (bg-pty-host) and 40175 (worker) were alive continuously from

2026-07-07T10:46:59Z (ps -o lstart) until manually kill -9'd 2026-07-08 ~07:20 CEST
-- roughly 20.5h running despite "done" status

  • daemon.log shows "bg adopt: adopted=1 respawned=0 dead=0" at both auto-upgrade

restarts (2026-07-07T21:11:00Z 2.1.202->2.1.203, 2026-07-08T05:04:38Z 2.1.203->2.1.204),
i.e. the daemon re-adopted this same job as live across two version upgrades

  • claude --resume on this session (or the FleetView "Needs input" row for it) hangs

with no further UI action possible

  • /stop, claude stop 779aeb30, and Ctrl+X (per the on-screen hint) did not visibly

clear it until pids were killed directly with kill -9 and ~/.claude/jobs/779aeb30/
was deleted manually

  • separately: ps default STARTED column shows bare "10:46AM" (no date) for anything

under 24h old, which reads as same-morning even a full day later -- worth noting
since it complicates anyone else diagnosing job age from ps output alone

bug_report_779aeb30.txt

Errors

[]

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗