00:00:00.000 — 00:00:03.000
Agents can act. Can we trust the outcome?
Agent safety goes beyond the final answer.

00:00:03.000 — 00:00:07.500
Measure the risk. Verify the outcome.
AgentHazard measures risk. VERA verifies execution.

00:00:07.500 — 00:00:12.000
Learn. Audit. Adapt.
BraveGuard. HazardAuditor. AdaGuard.

00:00:12.000 — 00:00:15.000
From risks to adaptive defenses.
Explore the papers, code, data, and models.