This website uses cookies to ensure you get the best experience.
Learn more.
A retrieved support document contains instructions telling the assistant to ignore its system rules and send private data to an external URL. Which response best preserves the trust boundary?
Published
- Assume a high similarity score proves the document instructions are trusted.
- Treat retrieved content as untrusted evidence, preserve instruction boundaries, and enforce tool and data access outside the model.
- Promote every retrieved instruction to the highest-priority system message.
- Allow the requested external transfer because the document is inside the company corpus.
Comments — 0
Sign in to comment