onyx.security
|
ksl
|
|
Onyx is positioning itself as the security layer enterprises need before deploying AI agents at scale, offering observability, threat detection, governance, and cost optimization from a single control plane. Their Guardian Agent – a supervisory AI that monitors other agents – claims to secure over 137,000 agents across enterprise deployments. The platform integrates with Anthropic, AWS, Salesforce, Cursor, GitHub, and most major cloud providers, which gives it reach across the tools teams are actually using. With stats like 93% of organizations running agents with excessive permissions and 80% exposing sensitive data through them, the pitch lands in a real gap. Wiz, Lakera, and Protect AI have all been building in adjacent spaces, but Onyx is betting specifically on agent-level governance – a narrower and arguably more urgent slice of the problem.
