AGENT 04 / 07
Operator
The reliable systems administrator. It schedules recurring workflows, manages durable background queues, keeps daemons healthy, and routes executive notifications.
operator · autonomous
01What it does
Unattended automation & scheduling
Durable job execution
Manages long-running background tasks with retry policies, state reconstruction, and restart survival.
Cron-like scheduling
Triggers periodic coworker workflows on precise cadences (watch mode, nightly inbox sweeps).
Multi-channel dispatch
Routes alerts and reports to webhooks, email SMTP, Slack, and Telegram channels.
Daemon process control
Starts, stops, monitors, and restarts background companion daemons with health probes.
Atomic schedule claiming
Coordinates multiple server worker instances using SQLite atomic state locks.
Operational telemetry
Exposes Prometheus metrics and structured runtime tracing for enterprise monitoring.
02How it works
From scheduled trigger to verified execution
Job submission
Enqueues background jobs with structured parameter schemas and retry policies.
Schedule evaluation
Evaluates cron triggers and claims tasks with atomic lock protection.
Execution & heartbeat
Monitors worker health, emits progress heartbeats, and tracks memory usage.
Error diagnosis
On failure, inspects error telemetry and applies intelligent self-healing retry strategies.
Alert dispatch
Delivers formatted PDF/HTML reports to configured team communication channels.
03Under the hood
The runtime underneath
| durability | Durable background job queue persisted in SQLite with state replay |
| scheduler | Cron-syntax scheduler with microsecond precision and atomic claim locks |
| telemetry | Prometheus metrics (/metrics) and structured tracing spans |
| daemons | Child process manager with auto-restart and stdout regex health checks |
04Tools