Skip to content

AML.T0054: LLM Jailbreak / Guardrail Bypass — TTP Profile

Density:
THREAT TECHNIQUE & TTP DOSSIER

AML.T0054: LLM Jailbreak / Guardrail Bypass

Framework: MITRE ATLAS • Tactic: Defense Evasion (AML.TA0005)

ACTIVE

30-Day Trend: +12% • Frequency: 2 sightings
Technique Identifier:
AML.T0054
Primary Tactic:
Defense Evasion
Active Adversaries:
1 attributed threat groups
Active Campaigns:
0 intrusion sets

  • Canonical Technique ID: AML.T0054
  • Governing Taxonomy: MITRE ATLAS
  • Operational Tactic: Defense Evasion (AML.TA0005)
  • Observed Telemetry Velocity (30d): +12%
  • Current Operational Status: ACTIVE
  • Recorded Detection Sightings: 2

Adversaries utilize sophisticated prompt engineering, semantic deception, token manipulation, or automated optimization algorithms to disable model safety guardrails and elicit forbidden behaviors.


AdversaryHermes IDMotivationSophistication
CLOSEDQUORUM Threat ClusterACTOR-DEV-CLOSEDQUORUMfinancialadvanced

No standalone discrete campaign bound; continuously monitored in global telemetry.

Malware FamilyIdentifierClassificationActive State
CLOSEDQUORUM Autonomous AI C2MAL-CLOSEDQUORUMagentic_implantACTIVE


4. Detection Engineering & Defensive Controls

Section titled “4. Detection Engineering & Defensive Controls”

Analyze prompt tokens for anomalous token sequences, high perplexity prompts, foreign language encoding tricks, and repeated refusal-triggering inputs.

Implement multi-stage moderation pipelines (input classifier + secondary validator LLM); enforce deterministic tool execution boundaries independent of model text responses.


5. Correlated Observables & Telemetry Indicators

Section titled “5. Correlated Observables & Telemetry Indicators”
Indicator TypeValueFirst SeenStatusSource
sha256e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 (IOC-2026-0002)2026-09-18ACTIVESRC-CISCO-TALOS