Claude Security Enters Public Beta for Enterprise Customers.

Anthropic has opened public beta access to Claude Security, a code scanning tool that puts the company's Opus 4.7 model to work on vulnerability detection without requiring custom infrastructure. The feature is now available to all Claude Enterprise customers.

The tool scans codebases for security vulnerabilities, validates findings to cut false positives, and suggests patches that development teams can review before deployment. Anthropic positions it as a turnkey alternative to building custom agent workflows or managing API integrations, a recurring request from security teams evaluating how to deploy frontier models in production.

Claude Security launched as a research preview in February. Since then, hundreds of organizations have run it against production code, according to Anthropic's announcement on LinkedIn. The company reports that early adopters identified vulnerabilities their existing scanners had missed, though it did not release specific numbers on detection rates or false positive reduction.

The public beta adds four features based on feedback from the preview period:

Scheduled scans that run automatically on a defined cadence, removing the need for manual initiation. Directory-level targeting to focus scans on high-risk areas of a codebase rather than scanning everything every time. CSV and Markdown exports to pipe results into existing security workflows, ticketing systems, or compliance dashboards. Documented dismissals that persist across scans, so triage decisions made during one scan carry forward and don't reappear as noise in subsequent runs.

The dismissal feature addresses a common friction point in vulnerability management: repeatedly seeing and re-evaluating the same low-priority findings. By documenting why a team dismissed a vulnerability, Claude Security can suppress it in future scans unless the underlying code changes.

Anthropic's pitch centers on accessibility. Security teams can activate Claude Security through the Enterprise interface without writing code, standing up infrastructure, or managing prompt chains. That contrasts with the growing number of organizations building custom security agents on top of Claude's API, which Anthropic acknowledged as "one path" in its announcement.

The distinction matters for resource allocation. Building a reliable agent workflow requires engineering time, ongoing maintenance, and expertise in both security and LLM orchestration. Claude Security packages those capabilities into a managed product, trading flexibility for speed to deployment.

That tradeoff will appeal most to teams that need coverage now and lack the bandwidth to build custom tooling, or to organizations that want to evaluate LLM-based scanning before committing to a larger integration project. Teams with complex requirements or non-standard workflows may still opt for the API route.

Claude Security remains exclusive to Enterprise customers, which require a minimum annual contract. Anthropic has not announced plans to bring the feature to Team or Pro tiers, though the company has recently expanded other Enterprise features to lower-tier subscribers after an initial rollout period.

The tool joins a growing list of domain-specific features Anthropic has shipped for Enterprise users in recent months, including HIPAA-compliant healthcare connectors and integrations for child welfare case management.

Enterprise customers can access Claude Security through their existing dashboards starting today. Anthropic has published documentation at the link in its LinkedIn post, though the landing page was not publicly accessible at the time of publication.