- •Building Caracal, an AI-agent security platform that enforces pre-execution authority checks before humans, agents, or services can trigger delegated actions
- •Designing policy-driven AI execution controls for request routing, approval gates, revocation checks, and common human-to-agent and agent-to-service workflows
- •Modeling delegation graphs that evaluate long and deep authority chains, resolve valid permission paths, and block unauthorized agent actions before execution
- •Integrating a secret vault layer so AI agents can use API keys and service credentials through policy-approved access without exposing raw secrets
- •Recording tamper-evident audit records and immutable proofs to a ledger-backed adapter for compliance, traceability, and AI-action accountability
- •Extending the platform toward agent policy negotiation, security reasoning modules, decentralized identity, verifiable credentials, and audit-derived policy recommendations
AI SecurityAI AgentsPolicy EnginesDelegation GraphsSecret VaultsDLTAudit LogsDID/VCAgent Security