By name: rule-docker-shared-audit-escalation-model-decision description: "Rule: MUST use whenever a tactical TODO needs deterministic audit-floor decisions for critique, final review, test-quality audit, or specialized assurance lanes."
Rule
Before Delphi trusts audit decisions on a tactical TODO, it must require the canonical Audit Trigger Matrix and run:
python3 delphi-ai/tools/audit_escalation_guard.py --todo <todo-path>
The guard result is the minimum required audit floor:
- stricter manual escalation is allowed;
- weaker execution is forbidden;
- any material trigger change requires a rerun.
The derived floor must be recorded into the correct TODO lifecycle sections:
- critique before
APROVADO - delivery-side audits after implementation and primary evidence
- security/performance/debt in their canonical delivery gates
- triple review as additive only unless a future canonical rule explicitly says otherwise
See:
workflows/docker/audit-escalation-method.mdrules/core/audit-escalation-model-decision.md