AI-based offensive cybersecurity firm Theori said on Oct. 16 it will join Anthropic's "Cyber Verification Program" (CVP).

The company said Anthropic's CVP grants vetted security companies and experts access to advanced cyber security AI capabilities that are typically restricted. It said the program allows the use of the latest AI models for high-level security research, including malware analysis, vulnerability detection and attack-path verification.

Theori said joining Anthropic's CVP will allow it to enhance the offensive security detection and attack-path verification engine of its AI-based autonomous penetration testing platform, Xint.

It said it expects to reduce false positives and speed up the discovery of high-risk vulnerabilities by adding AI agent reasoning capabilities in source code and binary analysis, as well as penetration testing processes in live service environments.

Earlier, Theori received approval to take part in OpenAI's cyber defense program, "Daybreak."

Theori CEO Se-jun Park (박세준) said, "As AI is rapidly lowering the time and cost required for cyberattacks, defenders must also be able to actively use the most advanced AI technologies." He said, "This approval is an important opportunity to use world-class AI technology for broader security research." He added, "Based on this, Theori will advance technology to find and verify real exploitable vulnerabilities faster and more accurately, helping customers respond to threats before attackers."

Theori provides its AI-based autonomous penetration testing platform, Xint.

The company said Xint consists of "Xint Code," an AI-based white-box penetration testing tool that tracks source code context and data flows to identify exploitable flaws in advance; "Xint Web," an AI-based black-box penetration testing tool that checks services in operation from an attacker's perspective; and "Xint Pulse," an on-demand penetration testing solution that checks individual web applications when needed.