Pylon is approved for Defense Access,
on Anthropic's Cyber Verification Program.
Vulnerability triage, malware analysis, detection engineering, and incident response, on systems we own or maintain.
Anthropic's announcement →Authorized penetration testing and red teaming. A separate application, which we have not made.
The Pylon Group has been approved for Defense Access under Anthropic's expanded Cyber Verification Program. Our team now has access to Anthropic's most capable cybersecurity models, including Claude Mythos 5.1, Claude Opus 5.5, and Claude Sonnet 5.5, for defensive security work.
What Defense Access covers
Defense Access is for teams defending systems they own or maintain. Within that scope, we will use these models for vulnerability triage and verification, malware analysis and reverse engineering, detection engineering, and incident response.
What it doesn't cover
Defense Access does not include authorized penetration testing or red teaming. Those activities fall under Red Team Access, which requires a separate application. We have not applied for it, and we will not describe ourselves as offering offensive testing until we have.
Why this matters for our work
Pylon is built AI-native. The models doing security work are the ones we would want defending our own systems, and access to the most capable models for defensive use is part of how we deliver that.
We will use this access inside the boundaries Anthropic sets for the program, including its guidance on agent containment.
The Pylon Group is an advisory and software studio for cybersecurity, technology, and AI, led by a practicing CISO and CIO.
We put senior operators into the seat, and we build the tools we wished existed when we ran those programs ourselves.
- Enquiries
- [email protected]
- Brand assets
- Logos and marks →
- Follow
- LinkedIn →