Claude has put Claude Security into public beta for Claude Enterprise customers, with the company saying the tool scans a codebase for vulnerabilities, validates findings to reduce false positives, and suggests patches for review and approval. The approach is similar to Cursor Security Review adds always-on agents for teams, although Claude’s rollout is aimed at security teams that want to use Opus 4.7 on code without building custom tooling. Claude states that no API integration or agent build is required.
In posts on X, the company adds that since a research preview in February, "hundreds of organizations" have used the product on production code and found issues that "existing scanners had missed." Claude also lists several additions based on early feedback: scheduled scans, directory-level targeting, CSV and Markdown exports, webhook notifications for new findings, and dismissals that carry forward across scans.
The rollout drew a mix of curiosity and skepticism in replies. Some users described the feature as a possible threat to existing security tools, while others questioned whether a "public beta" limited to Enterprise customers fits the label or differs much from running targeted audits in Claude Code.
Source: Claude on X
