Anthropic has released a new security plugin for its Claude Code development tool, designed to automatically detect and remediate vulnerabilities during the coding process. The update targets developers who want to integrate security checks directly into their workflow without adding manual steps or complex configuration.
Automated vulnerability detection runs during coding sessions without manual setup
The 'security-guidance plugin' operates as an official add-on within the Claude Code environment. It runs continuously alongside standard editing sessions, scanning code changes in real time rather than waiting for a separate audit phase.
The system uses a three-stage review process to identify potential security issues. The first stage applies fast pattern matching while files are being edited. A second background check occurs at the end of each conversation turn. A final detailed agent review runs when code is committed, aiming to catch injection flaws or unsafe deserialization before pull requests are submitted.
Installation requires no special commands or manual setup; the plugin activates automatically upon installation within the marketplace. This design allows developers to start using security checks immediately without learning new interfaces or altering their existing development habits.
Internal testing at Anthropic shows a 30-40% reduction in security-related comments during code reviews. The company reports that the tool helps catch common vulnerabilities earlier, reducing the need for back-and-forth corrections after initial drafts are submitted.
The plugin is available globally to all Claude Code users starting May 27, 2024. It can be installed directly from the official plugin marketplace without additional fees or regional restrictions.
Discussion
0 comments
Log in to join the thread with a thoughtful take, question, or correction.