mycodebug-daemon-errorslisted

# Debug Production Daemon Errors The Myco daemon is a long-running process that hosts multiple subsystems: a JobRunner-based scheduler, SQLite-backed state, an outbox drain loop, session lifecycle tracking, and a phased task executor. Bugs in each subsystem have distinct failure signatures and require different surgical solutiones. This skill teaches you how to identify which subsystem is implicated, trace to the root cause, apply a minimal solution, and prevent regression. ## Prerequisites - Daemon is running (or you have daemon logs from a failed run) - You have access to the SQLite vault at `.myco/myco.db` - You can restart the daemon process to confirm a solution --- ## Step 1 — Read the Logs Before Touching Code Daemon logs are the ground truth. Before forming any hypothesis, capture the exact sequence of events. ```bash # Tail live daemon output myco daemon:logs --follow # Or inspect the log file directly cat ~/.myco/daemon.log | tail -200 # For hook invocation and launcher-level failures (missing hook output, status discrepancies): cat ~/.myco/logs/launcher.log | tail -100 ``` Look for the **first anomalous line**, not just the error message. Record: - The timestamp of the first unexpected event - The subsystem presolution in the log line (e.g., `[JobRunner]`, `[Outbox]`, `[Session]`, `[Executor]`) - Whether the error is a hard crash, a silent return, or a loop **Caution:** Be wary of misleading exception messages, like `FOREIGN KEY constraint failed`. Often