Remote triggers silently fail to execute — next_run_at never advances

Summary

Remote triggers return HTTP 200 on both scheduled and manual runs but never actually execute. The next_run_at timestamp does not advance after a scheduled fire time passes, and no agent output is produced.

Steps to Reproduce

1. Create a remote trigger via the API with a daily cron schedule
2. Wait for the scheduled next_run_at time to pass
3. Observe that next_run_at remains frozen at the original scheduled time
4. Call the /run endpoint manually — returns HTTP 200 but produces no output

Expected Behavior

• Trigger executes at the scheduled time
• next_run_at advances to the next scheduled time after a run
• Manual /run call executes the agent and produces output (e.g., emails sent)

Actual Behavior

• No execution occurs (no emails sent, no agent activity observed)
• next_run_at remains frozen
• Manual /run returns HTTP 200 with no effect

Environment Details

• Trigger ID: trig_01BYVS3VHx6mUJ4dRRiTtcZS
• Environment ID: env_018nUtbgL7UPYGhU2PQLvGCT
• Cron expression: 0 10 * * *
• MCP connector: Gmail (connector_uuid: 5e6343f1-fd13-491f-9e5f-097e168e4cb1)
• Model: claude-sonnet-4-6
• Account: alex@taube (alexander.j.taube@gmail.com)

Additional Notes

A second trigger (trig_01RHWSEumuWMxDcki7P8663R, now disabled) on the same account exhibits the same behavior — it was created April 3, 2026 and also never executed successfully.

Recreating the trigger via API is not possible without a valid environment_id, which is not obtainable through the trigger API itself.