Banana Navy
Catalog FR
Lab · AI threat modeling · Detailed card

Slow degradation and slow attacks

Eroding the service in small steps, with no spike to raise the alarm.

CardF12
CategoryD-TB7-01 (gradual disruption)
Layers19 · Monitoring
SystemAI voicebot

Watch slow variations and bound the degradation.

The threat

slow probing, stretched campaigns, progressive service degradation up to partial unavailability, with no spike that trips classic alarms.

Blind spotWhy classic frameworks miss it

alert thresholds are built on spikes; an under-threshold attack stays invisible while fragmenting the signals.

MitigationProposed approach

telemetry (OpenTelemetry, Grafana), anomaly detection, alerts on slow variations, bounded capacity and degradation.

The proposed control
telemetry and degradation bounds on the service.

Expected evidence
a slow attack is detected before operational impact.

SourcesReferences and public research

OWASP GenAIOWASP GenAI LLM10:2025 Unbounded Consumption
Public researchPublic research sources: MITRE ATLAS 2026.07 (verified technique mapping), OWASP GenAI, and the public risk-voicebot (aivansoul/risk-voicebot) template defining the 20 checkpoints. No client registry data: generic card, no rating, no verdict.

Explore the 20 security layers

MITRE ATLAS 2026.07 · OWASP GenAI · risk-voicebot