SCBE Operational Runbook
Last updated: 2026-02-17
Goal
Run SCBE-AETHERMOORE from kernel validation through browser execution, orchestration, cloud deployment, and training data workflows in a deterministic order.
Step 1: Validate canonical contracts
Commands:
python scripts/validate_layer_manifest.py
Checks:
docs/scbe_full_system_layer_manifest.jsonformat and required fields.- Layer dependency references (
depends_on) are valid. last_verified_commitvalues are structurally valid hashes.- PR CI enforces non-downgrade status transitions via
scripts/check_layer_status_transition.py.
Step 2: Bring up local governed browser service
Commands:
.\scripts\run_aetherbrowse_service.ps1 -SCBEKey "<32-char-key>" -Port 8001 -KillOnPortInUse
Health check (new terminal):
Invoke-RestMethod -Method Get -Uri "http://127.0.0.1:8001/health"
Step 3: Execute multi-agent browser jobs
Commands:
python scripts\aetherbrowse_swarm_runner.py --tasks examples\aetherbrowse_tasks.sample.json
Expected artifact surfaces:
training/runs/output payloads.- Decision-record linked results per job.
Step 4: Execute orchestration connectors (Asana + n8n)
Commands:
python scripts\asana_aetherbrowse_orchestrator.py --help
python scripts\n8n_aetherbrowse_bridge.py --help
Required environment:
ASANA_TOKENN8N_API_KEYSCBE_API_KEYSCBE_BROWSER_WEBHOOK_URL
Step 5: Deploy cloud execution plane (Cloud Run)
Commands:
bash deploy/gcloud/deploy_aetherbrowse.sh
Required environment:
GCP_PROJECT_IDGCP_REGIONSCBE_API_KEY
Step 6: Run training/data plane
Commands:
python scripts\notion_access_check.py
python scripts\notion_to_dataset.py
python scripts\push_to_hf.py
Required environment:
NOTION_API_KEYHF_TOKENGCP_SA_KEY(for Vertex workflows)
Step 7: CI gates and release handoff
Commands:
gh workflow run "Validate Layer Manifest"
gh workflow run "ci.yml"
Release condition:
- Layer manifest validator passes.
- Core CI gates pass.
- Any layer status changes are reflected in
docs/scbe_full_system_layer_manifest.json.