
CUAHarm
Benchmark for evaluating safety risks of computer-using agents, with 104 realistic misuse scenarios across seven malicious categories, supporting…
ai-securityeducationlabs-practice+3
1

Benchmark for evaluating safety risks of computer-using agents, with 104 realistic misuse scenarios across seven malicious categories, supporting…

Iterative agent harness that uses LLMs and Certora Prover to generate and refine smart-contract CVL specs, feeding verifier output back until success…