Darkhunt AI is building the security control plane for AI agents. We continuously discover and validate real attack paths against agentic AI systems.  Our approach uses swarms of AI agents to quickly and accurately explore attack scenarios, identify risky behaviours, and surface actionable fixes, which allows teams to ship AI products faster and with confidence.

The key focus is closing the loop: automatically discovering and learning from new attack techniques, then converting them into protections - defences and guardrails that continuously harden AI systems over time without disrupting engineering workflows

Darkhunt AI

Org chart

Teams

Offices