Evolved agent safety harnesses cut adaptive injection success
On AgentDyn with GPT-4o-mini, attack success fell from 50.4% to 3.0% while utility improved.
2 stories, newest first · one card per story
On AgentDyn with GPT-4o-mini, attack success fell from 50.4% to 3.0% while utility improved.
EvoSafeHarness auto-synthesizes per-model, per-domain safety harnesses, cutting prompt-injection attack success on AgentDojo to 0.0% at 82.8% utility.
Today 2.5K readers · 2.7K views · 144 reads · all time 206K views