L14Confidence ScoringArt. 14Decision confidence, human review routing
L13MCP Server RegistryArt. 15Supply chain validation for AI tool servers
L12Behavioural IntelligenceArt. 9, 72Baselines, anomalies, multi-agent correlation
L11Network SecurityArt. 15Mutual auth, replay protection
L10Data ProtectionArt. 10PII, credentials, egress, classification
L9Meta-SecurityAnnex VISelf-protection of the governance platform
L8Tool SandboxArt. 15Isolated execution, signed manifests
L7Memory IntegrityArt. 15Cryptographic hashing, tamper detection
L6Input SanitiserArt. 1530+ prompt injection/jailbreak patterns
L5Model Quality GateArt. 155-tier certification for AI models
L4Agent PoliceArt. 9, 72Real-time monitoring, risk scoring, quarantine
L3Compliance EngineArt. 14Rules, spending limits, approval queues
L2Agent Instruction LanguageArt. 15Encrypted inter-agent communication
L1Agent Identity ProtocolArt. 13Unique identity, certificates, lifecycle