ClawHosters
by Daniel Samer
3 min read
The parent company behind Norton and Avast just built a free security layer for OpenClaw. That alone should tell you how far the ecosystem has come.
On March 26, 2026, Gen Digital and the OpenClaw security team co-hosted a post-RSA event in San Francisco's Financial District. The format was a fireside chat, a builder panel, and a hands-on preview of what they're calling the Agent Trust Hub (ATH).
What the Agent Trust Hub Actually Does
Two things worth paying attention to here.
First, the AI Skills Scanner. It's a free diagnostic tool that checks OpenClaw skills for malicious patterns before you install them. Think of it as a virus scan, but for agent behavior. Second, a curated AI Skills Marketplace where every listed skill has been audited.
Both are free at ai.gendigital.com/agent-trust-hub.
Howie Xu, Gen's Chief AI & Innovation Officer, put it plainly during the event: "AI agents are moving quickly from concept to real-world action, making security and trust critical."
He's not wrong. Gen Threat Labs found over 18,000 exposed OpenClaw instances online. Roughly 15% of the skills they reviewed contained malicious instructions. That's not a theoretical risk. That's a real problem affecting real deployments right now.
Why This Matters for Self-Hosters
Siggi Stefnisson, Gen's Cyber Safety CTO, made a point that stuck with us. Security failures aren't just about one bad click anymore. A compromised skill can turn a trusted AI assistant into a persistent insider threat. Quietly. Without the user noticing.
If you're self-hosting OpenClaw, the Skills Scanner gives you a way to vet skills before they touch your instance. It's probably the single most useful free tool in the OpenClaw security space right now. For anyone running a managed ClawHosters instance, we already handle security hardening on the infrastructure side. But skill-level trust? That's been a gap until now.
The Bigger Picture
Gen is the first major consumer cybersecurity company to build a dedicated trust layer specifically for OpenClaw. They're joining Cisco DefenseClaw, Trend Micro TrendAI, and NVIDIA NemoClaw in what's becoming a real security ecosystem around AI agents at RSAC 2026.
For a deeper look at how OpenClaw security works on the infrastructure level, check our security hardening guide. And if you want to get started with a secured instance, sign up here.
