Claude Security is now monitoring software repositories as code changes, giving enterprise development teams a new way to find and manage security issues throughout the coding process.
The service is available in public beta for Claude Enterprise customers. It tracks findings until they are fixed, uses context supplied by teams, and connects with existing tools through an API and the Model Context Protocol, known as MCP.
Security Checks Follow Code Changes
Traditional security reviews often happen at fixed points, such as before a release. Continuous repository monitoring shifts that work closer to daily development.
Claude Security watches repositories as engineers change code. This approach may help teams identify problems earlier, when fixes are often easier and less costly to complete.
“Claude Security now watches your repositories as code changes, tracks findings until they’re fixed, learns from the context your team adds, and connects to your tools through an API and MCP.”
The announcement does not explain which programming languages, repository platforms, or vulnerability types are supported. It also does not publish data on detection rates, false positives, or scan speed.
Those details will matter to security leaders. Automated code analysis can create large numbers of alerts, and developers may ignore them if too many lack clear evidence or practical fixes.
Context Could Improve Alert Quality
The ability to learn from team context is a central feature. Security findings often depend on how an application is built, deployed, and protected.
For example, code that appears exposed during an isolated scan may sit behind access controls. In another case, code that looks safe may handle sensitive data under strict internal rules.
Team-supplied context could help Claude Security account for such differences. However, organizations will need clear policies for the information they provide and how that information affects later findings.
Key questions for beta users include:
- How accurately the system ranks urgent findings.
- Whether it explains why code creates a security risk.
- How it handles incorrect or outdated team context.
- What repository data is retained or processed.
API and MCP Support Broaden Integration
API access allows companies to connect Claude Security with internal workflows. Possible uses include sending findings to ticketing systems, development dashboards, or security review processes.
MCP offers another route for connecting Claude with tools and data sources. For enterprise customers, these links may reduce manual handoffs between developers and security teams.
Integration also creates governance concerns. Companies will need to control permissions, record automated actions, and decide who can dismiss or change a finding. Human review will remain important, especially for issues affecting production systems or customer data.
Public Beta Will Test Enterprise Value
The public beta gives Claude Enterprise customers an opportunity to assess the service under real development conditions. Feedback may shape its accuracy, workflow support, and administrative controls before wider availability.
No general release date, pricing details, or availability for non-enterprise users was announced. The initial scope suggests a focus on organizations with established security teams and formal software delivery processes.
Claude Security’s value will depend on more than finding possible flaws. Enterprises will watch whether it reduces unresolved alerts, fits existing tools, and helps developers fix issues without slowing releases. The beta’s strongest measure will be whether tracked findings lead to verified fixes.
