Overnight review

Overnight review is the deeper process window for quality and drift detection. It complements daytime integrity sweeps.

Window

• Recommended local window: 18:00-06:00.

Goals

• Catch regressions that only appear over longer runtime windows.
• Distinguish product defects from auth/env/testbench drift.
• Create durable context updates for future OpenClaw clones.

Expected outputs

• Environment mode and mode basis.
• High-signal changes in BACnet behavior, graph behavior, and fault outputs.
• Candidate docs/playbook updates when operating process improves.
• If durable bench/operator lessons emerge, capture them in versioned notes under openclaw/references/ so they can later be curated into broader docs.

Alerting discipline

• Suppress repeated low-signal notices.
• Escalate only meaningful drift or new high-impact findings.
• Convert durable lessons into versioned docs, not chat-only memory.