Broader Access and Key Capabilities
Anthropic has opened Claude Security, its AI powered vulnerability detection tool, to a wider audience of enterprise customers as a public beta. The tool, built on the Opus 4.7 model, is designed to identify and fix software flaws. It can explain a vulnerability in detail, rank its confidence in the severity and real world impact, and generate step by step instructions for a targeted patch. Users can open these instructions in Claude Code on the web to work through the fix directly.
Enterprise Integration and Guardrails
Several major cybersecurity firms, including CrowdStrike, Microsoft Security, Palo Alto Networks, SentinelOne, TrendAI, and Wiz, are embedding Opus 4.7 into their own products. Palo Alto Networks stated the integration helps customers outpace automated threats. For users who have not verified themselves as cybersecurity professionals, Anthropic has put guardrails in place to limit misuse. The new iteration also includes a multi stage validation process to reduce false positives, an option to schedule scans, and the ability to scan specific directories.
Market Skepticism and Expert Views
Despite the buzz around Claude Security, some cybersecurity veterans remain cautious. Entrepreneur Jeremiah Grossman noted that most CVEs never get weaponized in real world attacks, arguing that increased capability does not change what attackers actually need. Kevin Beaumont echoed this sentiment, criticizing the panic around generative AI as driven by people without real world context. Anthropic also restricted access to its more powerful Mythos Preview model, which remains limited to about 50 companies under Project Glasswing due to national security risks.
Source: Healthcareinfosecurity
