fix(status): preserve Patrol Fix progress across restarts

[ivankuznetsov]

Summary

After a daemon restart, operational status no longer turns a receipt-progressed Patrol Fix task into needs_repair only because a completed predecessor attempt left a dead PID in .lock. Durable workflow progress stays scheduler-owned, and the next dispatch reclaims the stale lock automatically. Current agent_running, AGENT_WORKING, and REVIEW_WORKING claims with a dead recorded runner still report repair.

Validation

• bundle exec ruby -Ilib -Itest test/unit/operational_status_test.rb (59 runs, 275 assertions)
• bundle exec ruby -Ilib -Itest test/unit/commands/status_operational_test.rb (6 runs, 18 assertions)
• bundle exec ruby -Ilib -Itest test/unit/schema_files_test.rb (161 runs, 910 assertions)
• bundle exec ruby -Ilib -Itest test/unit/lock_test.rb --name test_stale_lock_with_dead_pid_is_replaced (1 run, 1 assertion)
• git diff --check

bundle exec rubocop could not start because the host mise configuration has no Ruby version selected for the rubocop shim. Hosted CI remains the lint gate.

---

[ivankuznetsov]

hive-babysitter could not repair this PR automatically.

Outcome: failure

Agent tail:

You've hit your session limit · resets 7:40pm (Europe/London)