Dokumentationerreichbar
Anthropic, Catch security issues as Claude writes code
code.claude.com (externe Seite)
Die Dokumentation zum Plugin security-guidance. Claude prüft damit die eigenen Änderungen auf Sicherheitslücken und behebt sie in derselben Sitzung.
geprüft 10.11.2026
Worauf sich diese Seite beruft, wörtlich, abgerufen am 12.10.2026:
The plugin does not ask the same Claude instance that wrote the code to grade itself.
bestätigt 10.11.2026the reviewer starts from the diff, has no investment in the original approach, and is instructed only to find problems
bestätigt 10.11.2026The security guidance plugin makes Claude review its own code changes for common vulnerabilities while it works and fix what it finds in the same session.
bestätigt 10.11.2026The plugin reviews Claude’s work at three points, each at a different depth
bestätigt 10.11.2026The plugin is available on all plans.
bestätigt 10.11.2026None of the layers block writes or commits.
bestätigt 10.11.2026The end-of-turn and commit reviews each spend additional model usage that counts toward your usage like any other Claude request.
bestätigt 10.11.2026The plugin is built entirely on hooks
bestätigt 10.11.2026Commits you run from your own shell, including the ! shell escape inside a session, are not reviewed.
bestätigt 10.11.2026For hard enforcement, pair the plugin with a hook that blocks the edit or a CI check.
bestätigt 10.11.2026By default, both reviews run on a Claude Opus model.
bestätigt 10.11.2026The YAML forms require Py
bestätigt 10.11.2026YAML to be importable, which the plugin does not install for you.