Threat Model

AegisGate’s threat model is structured using the STRIDE framework and maps to the MITRE ATLAS adversarial tactics matrix.

Attack Surface

SurfaceProtocolPrimary Threats
AI prompt inputHTTP, MCPPrompt injection, jailbreak, data exfiltration
AI response outputRESPONSEToxic content, PII leakage, instruction echo
Inter-agent communicationA2A, ACP, ANPAgent impersonation, task manipulation, context poisoning
Management APIHTTPUnauthorized config changes, credential theft
Supply chainAllDependency vulnerabilities, model poisoning

STRIDE Analysis

Threat TypeRiskMitigation
SpoofingAgent identity forgeryCapability tokens, HMAC verification
TamperingPrompt/response manipulationIntegrity checks, ML detection
RepudiationDenial of security eventsImmutable audit log, ATLAS mapping
Information disclosurePII leakage in responsesDLP patterns, PII redaction
Denial of serviceToken exhaustion, API floodingRate limiting, token budgets
Elevation of privilegeAgent privilege escalationRBAC, capability-based access

ML-Powered Detection

v4.2.0 introduces a CNN-BiLSTM neural network for prompt classification:

  • 100% detection rate on adversarial test suite (100/100)
  • 0% false positive rate on benign inputs
  • ~6ms inference latency (ONNX Runtime)
  • Graceful degradation to regex-only if ONNX unavailable

See also: Security Overview and MITRE ATLAS Coverage.