A proposed microVM agent runtime brokers outbound connections and credentials through host policy, while documenting residual trusted-destination risk.
Setting: research demonstrationExploitation: not establishedEvidence: Grade CRemediation: not applicable
Authors tested 44 reproducible flaws in 12 apps and found many attacks reachable through AI-mediated interfaces, with different tool interfaces exposing different capabilities.
Setting: evaluationExploitation: demonstratedEvidence: Grade CRemediation: not applicable
A verification system checks structured application state and selectively abstains, improving correctness on tested GUI-agent benchmarks at added cost.
Setting: evaluationExploitation: demonstratedEvidence: Grade CRemediation: not applicable
Talos describes malware that collects credentials and passes victim profiles to model APIs for prioritization, without establishing a count of compromised victims.
Setting: research demonstrationExploitation: not establishedEvidence: Grade BRemediation: mitigation available