Agent privacy evaluation: use AgentDojo to turn 'an agent leaking private data while doing its job' into a measurable benchmark
In one sentence it turns "an agent doing legit work while leaking the private data it sees" from "we feel it's probably safe" into a pre-release, regression-able privacy eval — measure the injection-driven exfiltration success rate (ASR) for each outbound tool and fold it into the release gate.