NEAR Protocol (@NEARProtocol): NEAR AI and FailSafe launch AttackBench: an open-source benchmark for real-world AI agent security.
Estimated coin impact
Scores are not price-change percentages. High impact means a score of +70 or more, or -70 or less.
- NEAR+30Positive
NEAR AI and FailSafe launch AttackBench: an open-source benchmark for real-world AI agent security. Real attackers adapt. Security standards should too. The future of agentic systems depends on verifiable resilience, not checklists.👇