New Archestra's OpenAPPA Saturates Two Major Security Benchmarks with a 0% Attack Success Rate
Archestra has released OpenAPPA, an open-source security engine built to prevent data exfiltration resulting from prompt injection or model hallucinations. In tests using the Bench-Corp and AgentThreatBench benchmarks, the system successfully blocked all malicious attempts, compared to a ten percent success rate for attacks against other models. This tool aims to provide a standardized defense layer for enterprise AI workflows by automating the detection and mitigation of input-based threats.
Covered by 1 source
- IInfoQ AI↗Bruno Couriol2d ago