Workers и автоматизирана поддръжка

Какво обхваща

• Runtime loop/local agent обновяват local state/heartbeat; apply orchestrator/queue изпълняват approved operations; recovery coordinator handles incomplete operations; event journal retains events.
• Hardware worker/scheduler събират catalog/evidence; permission baseline/integrity worker detects drift; supervisor/watchdog monitor services/workers.
• Worker finding трябва да има generated time, source, evidence, severity, subject и recommended plan.

Как се работи

1. Open finding.
2. Check freshness/evidence.
3. Check dependency/load.
4. Review recommendation.
5. Approve/reject controlled action.

Важни правила

• Worker не executes arbitrary shell.
• Не masks failure as success.
• Не deletes evidence/audit.
• Не blocks boot indefinitely.
• Не auto-accepts high risk.