Skip to content

Secure your app: mitigate risks to agentic features

Explore how to evaluate threats from indirect prompt injection, such as data exfiltration and unintended actions. Discover system safeguards and security best practices for using App Intents and the Foundation Models framework, including mitigations such as user confirmations, secure prompt design, and authentication.

✍️

No notes available yet

Be the hero who changes that. Watch the video, jot down what matters, and open a pull request – it's genuinely quick.

Learn how to contribute