Capture Notes
Paper evaluating data leakage risks in realistic tool-using LLM agent scenarios.
AI security relevance:
- Directly relevant to agent tool privileges, data exfiltration, and enterprise agent deployments.
- Good candidate for high-priority ingest.