메뉴
Generate or refresh the executive risk register spreadsheet aggregating all findings with severity scores, owners, and remediation priorities.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Trace a public endpoint through App Gateway, APIM, and AKS using harvested subscription data.
On-demand Playwright audit of a subscription's rendered React Flow diagrams - verifies the UI matches harvested Azure reality using screenshots, DOM checks, interaction tests, and diagram-to-asset comparison.
Run Triage-Saurus diagram review (screenshots, threat-model smells, and before/after report).
Promote validated learnings from a completed experiment — create detection rules, update strategy JSON, and update agent instructions for future experiments.
Validate generated architecture diagrams for hierarchy issues, missing internet ingress, network segmentation gaps, and missing components.
Generate per-resource-type cloud summaries (Key Vault, Storage Account, AKS, etc.) and persist them to the DB for the web app after triage completes.
| name | risk-register |
| description | Generate or refresh the executive risk register spreadsheet aggregating all findings with severity scores, owners, and remediation priorities. |
Generate or refresh the risk register after triage (or score changes) using findings from the DB.
Full agent guidance is in Agents/RiskRegisterAgent.md.
At least Phase 4 (enrichment) must be complete for the findings to have scores. Phase 5 (skeptic reviews) should ideally be complete so final scores are used.
Verify experiment state with:
python3 Scripts/Experiments/triage_experiment.py resume
python3 Scripts/Utils/risk_register.py
The script uses hardcoded output paths from Scripts/Utils/output_paths.py — no flags required.
python3 Scripts/Utils/watch_risk_register.py
Watches Output/Findings for score changes and regenerates the register automatically — useful while skeptic reviews are running.
| Flag | Purpose |
|---|---|
--findings-dir <path> | Directory to watch (default: Output/Findings) |
--interval <sec> | Poll interval in seconds |
--debounce <sec> | Debounce delay before regenerating |
--once | Run once and exit (no watch loop) |
Output/Summary/
Risk Register.xlsx ← executive spreadsheet (sortable by score, owner, resource)
Re-run the risk register after any of:
Score change: note added to finding)When a score changes, the reason must be documented in the finding file:
Score change: 5/10 ➜ 7/10 — confirmed internet-facing prod exposure
The risk register will reflect the latest score from the DB; the audit trail is in the finding MD.
Agents/RiskRegisterAgent.md