# MON-07 Agent-specific incident response

Domain: MON Continuous Monitoring & Human Oversight · HI-AAF v1.3 · Public review · https://humanintel.net/standard/mon/mon-07

An incident response plan specific to agent failures is in place, covering containment, investigation, customer communication, regulatory notification, and post-incident review.

Incidents involving AI agents are governed by a documented response process that covers containment, investigation, customer communication, regulatory notification, and post-incident review. The response plan is agent-specific: it accounts for the unique characteristics of agent failures, including the need to review agent reasoning traces, the possibility of cascading failures in multi-agent systems, and the challenge of determining root cause in non-deterministic systems.

## Maturity

- L1 Ad hoc: Controls exist informally; documentation is incomplete; no consistent review.
- L2 Documented: Controls are written down; ownership is assigned; review cadence is defined.
- L3 Operated: Controls are followed in practice; evidence is collected; deviations are tracked. Published criterion for MON-07: Incident response plan exists and covers agent-specific scenarios; the plan has been exercised in a drill or live incident; post-incident review records are retained; findings feed back into the evaluation set and policies.
- L4 Measured: Control effectiveness is measured against documented thresholds; trends are reported.
- L5 Continuously Improved: Controls evolve based on incident learning; maturity gains are evidenced over time.

## Evidence an assessor asks for

- Incident response plan exists and covers agent-specific scenarios.
- The plan has been exercised in a drill or live incident.
- Post-incident review records are retained.
- Findings feed back into the evaluation set and policies.

## Framework mappings

- NIST AI RMF: MANAGE 4.1, MANAGE 4.2
- SOC 2 (TSC): CC7.3, CC7.4, CC7.5
- EU AI Act: Art. 14(4), Art. 26(5)
- Singapore: MAIGF Incident Reporting

Mappings are directional: meeting this control evidences but does not by itself satisfy the external requirement.

## Related

- Cited by IAM-08 Multi-tenant credential isolation: https://humanintel.net/standard/iam/iam-08
- Cited by DAT-02 Cross-tenant data isolation: https://humanintel.net/standard/dat/dat-02
- Cited by DAT-11 Memory and retrieval isolation: https://humanintel.net/standard/dat/dat-11
- Cited by OUT-15 Model availability & continuity: https://humanintel.net/standard/out/out-15

## Cite this control

Human Intelligence. HI-AAF v1.3, MON-07 Agent-specific incident response. Public review edition, July 2026. https://humanintel.net/standard/mon/mon-07