Research

Watch · narrated whiteboard episodesL2

Agent Identity and Lifecycle Governance

An autonomous agent is a non-human principal that provisions itself, calls other services, delegates to sub-agents, and outlives any single request — and human identity systems were never designed for it. This series builds operational identity governance for agents from the ground up: why agents break human IAM and what a machine identity actually needs, how to provision and attest an identity so its bearer can be proven rather than assumed, how to run the credential lifecycle with short-lived tokens, rotation, and bounded delegation chains at runtime, how to decommission and revoke an agent's authority cleanly and provably, how to govern a whole fleet with an identity inventory, least-privilege drift detection, and lifecycle audit, and how to prove revocation actually works with recurring drills that time when every verifier, including partner services, starts refusing. Grounded in NIST SP 800-207 Zero Trust Architecture, NIST SP 800-63 Digital Identity Guidelines, the SPIFFE workload-identity specification, the OAuth 2.0 and Token Exchange RFCs, the OWASP Agentic Security Initiative, and the NIST AI RMF.

Murali Chillakuru·6 episodes
  1. 21 min Episode 1Non-Human Identity for Agents: Why Agents Break Human IAM and What a Machine Identity NeedsWhy identity systems built for people fail autonomous agents, assumption by assumption, and what a machine identity must provide instead: attributable, cheap to verify, short-lived, task-scoped and machine-governed.
  2. 22 min Episode 2Provisioning and Attestation: Issuing, Binding, and Proving an Agent's IdentityHow a brand-new agent gets its first credential without any pre-shared secret: why the shortcuts fail, attestation versus assertion, node then workload layers, the handshake, selectors judged by who controls them, binding through proof of possession, an attacker-by-access view, and the chain of trust.
  3. 21 min Episode 3Credential Lifecycle: Short-Lived Tokens, Rotation, and Delegation Chains at RuntimeWhat happens to an agent's credential after it is issued: why it should be born expiring, the trade-off at both ends of lifetime, the five token states, seamless rotation with overlapping windows, delegation by exchange, monotonic narrowing, and what attackers still have.
  4. 21 min Episode 4Decommissioning and Revocation: Killing an Agent's Authority Cleanly and ProvablyHow to end an agent's authority now and prove it is gone: why expiry is not revocation, what makes a revocation provable, four mechanisms and their latencies, propagation and stale verifiers, a latency budget, decommissioning the whole delegation tree, and choosing by urgency and reversibility.
  5. 21 min Episode 5Governing the Fleet: Identity Inventory, Least-Privilege Drift, and Lifecycle AuditFrom securing one agent identity to governing all of them: why fleet questions depend on completeness, continuous reconciliation of intended and actual state, an inventory complete by construction and modelled as a graph, least-privilege drift, lifecycle audit, and a maturity ladder climbed in order.
  6. 19 min Episode 6Revocation Drills: Measuring When Every Verifier Says NoHow to prove that cutting off an agent actually works: revoke a test agent on purpose, time when every verifier starts refusing, read the results in four outcome classes, deal with partner services you don't control, watch what the agent itself does, and keep evidence an auditor can rely on.