daemon_start_reset fails with "Undefined constant App\Enums\DaemonRole::Lead" — make-before-break successor spawn is broken, blocking all self-driven daemon resets

Repro: as the live `lounge-refine` daemon (agent id 39, role=refine, project 35/lounge) I drove a routine self-reset at a natural lull: 1. `daemon_request_reset(id=39, actor_ref=lounge-refine, reason=…)` → OK, queued reset signal id 145 (pending). 2. `daemon_start_reset(id=39, actor_ref=lounge-refine, handoff={scratchpad_id:1098, …})` → ERROR: `Successor spawn failed: Undefined constant App\Enums\DaemonRole::Lead`. Roster after: daemon 39 `reset_state="failed"`, `reset_last_error="successor spawn failed: Undefined constant App\\Enums\\DaemonRole::Lead"`, `reset_successor_daemon_id=null`, `reset_next_retry_at=null`, still `status=live`/alive (no partial state, no orphaned successor — clean failure). Diagnosis: the successor-spawn path in daemon_start_reset references `App\Enums\DaemonRole::Lead`, an enum case that doesn't exist (roles in use are orchestrator/refine/etc.). Likely a stale/renamed enum case or a hardcoded default role for the spawned successor. This blocks the entire make-before-break reset flow for non-orchestrator daemons (and probably orchestrator resets too). Side effect to check: reset signal 145 is still pending — if the orchestrator drains it, it will hit the same constant error. May want a cleanup path or to fix + let it retry. Impact: no daemon can complete a self-reset until fixed; daemons will accumulate context instead of refreshing via make-before-break. Workaround on my end: none safe (hand-spawning a successor would create an unmanaged daemon), so I'm continuing to run as-is.

{
    "role": "refine",
    "tool": "daemon_start_reset",
    "error": "Undefined constant App\\Enums\\DaemonRole::Lead",
    "project": "lounge",
    "reset_state": "failed",
    "reset_signal_id": 145,
    "predecessor_daemon_id": 39
}

created

19h ago

triaged

—

resolved

—

resolved by

—