Onyxia Accepted into Anthropic’s Cyber Verification Program
We're excited to announce that Onyxia Cyber has been accepted into Anthropic's Cyber Verification Program (CVP) — a designation reserved for security organizations using advanced AI capabilities strictly for defensive purposes.
For our customers, this isn't just a credentialing milestone. It directly expands what Nexa, our AI Security Agent, can do on your behalf.
Why This Matters
Today's security teams are stretched past their limits. The average enterprise runs more than 60 security tools, each generating its own stream of alerts, findings, and recommendations — most of which never connect to a coherent picture of actual risk. Remediation happens in silos. The same exposures resurface quarter after quarter. And as SEC regulations and global compliance mandates expand executive accountability, CISOs now face consequences for outcomes their existing stacks were never built to prevent.
Onyxia was built to change that. Our platform connects a cross-domain security data fabric with Nexa's agentic reasoning to move security teams from reactive response to preemptive action — surfacing exposures before they become incidents, and delivering remediation plans that are ready to execute.
Membership in Anthropic's CVP deepens that capability at every layer.
What the CVP Unlocks for Nexa
Anthropic's Cyber Verification Program grants approved organizations access to advanced AI capabilities that are normally restricted in standard model deployments — specifically for defensive use cases, including threat analysis, exposure modeling, security data integration, and risk mitigation.
For Onyxia, this means Nexa can now apply richer, more nuanced reasoning to the cross-domain security context it assembles for each customer environment. Where standard model guardrails previously limited certain analytical depth, CVP membership allows Nexa to:
Perform deeper threat and exposure analysis across fragmented data from disparate tools, connecting signals that would otherwise remain siloed
Generate more precise, actionable mitigation plans with greater confidence in the underlying risk assessment
Power more sophisticated security stack integrations, enabling Nexa to reason about how data flows across your environment and where exposures compound
The practical result: faster time to actionable insight, and fewer gaps between what your tools can see and what Nexa recommends you do about it.
Responsible AI in Security — By Design
Anthropic's CVP isn't a rubber stamp. Acceptance requires demonstrating that your organization's use of advanced AI capabilities is scoped, governed, and genuinely defensive in nature.
We're proud that Onyxia's approach to agentic security — centered on operational cyber resilience over reactive detection — aligns with exactly the kind of responsible, outcome-driven use Anthropic designed the program to support. Our customers aren't just getting more powerful AI. They're getting AI that's been vetted for how it operates in high-stakes security environments.
The Outcomes Speak for Themselves
CVP membership accelerates what our customers are already experiencing on the platform:
A healthcare organization reduced a multi-million-dollar regulatory fine to hundreds of thousands with the help of Onyxia’s historical, cross-domain program data
Multiple enterprises now save over 160 team hours per month on manual governance reporting
A retail manufacturing company identified and remediated exposures they didn't know existed — within a week of onboarding
These aren't outcomes driven by better dashboards. They're driven by Nexa connecting the dots across your security data and mobilizing the right response before regulators, auditors, or adversaries get there first.
What's Next
We'll continue expanding Nexa's capabilities as CVP membership deepens our access to frontier AI models and their defensive potential. For our customers, that means an agent that keeps getting better at turning complex, cross-domain security data into clear, prioritized action.
For CISOs and security leaders still managing risk through disconnected tools and manual reporting — we'd like to show you what proactive operational resilience looks like in practice.
