ChatML Role Injection (CRI)
blog.sentry.security | blog
Summary
Learn how ChatML Role Injection (CRI) attacks exploit JSON schema validation and how to secure your OpenAI agent against untrusted role injection
- Published
- Collected
Skip to content