AKAV LABS

अकवि · the unseeing seer

AI agents now outnumber human identities
inside enterprises by 80 to 1.

Every security tool was built for humans.

We're mapping the gap.

Subscribe to AgentSentry ResearchView detection rules
84

MITRE ATLAS techniques for attacking AI systems

0

Commercial products with detection rules for all 84

This is the gap we're filling.

atlas-agent-rules

MITRE ATLAS detection rules — all 84 techniques mapped

GitHub starsMIT license
View on GitHub

AgentSentry Research

Weekly AI agent security intelligence. One technique, one attack scenario, one detection rule. Every Tuesday.

Subscribe on Substack