CVE-Bench

Evals

Benchmark / dataset

Agents exploit real-world critical-severity web app CVEs in sandboxes.

Category
Cyber offense capability
Insurance exposure
Breach/cyber
Publisher
UIUC (Kang Lab)
Publisher type
Academic
License
Apache-2.0
Status
Active (<90d)
Last commit
2026-09-24
Stars
298
Notes
v2.1 Jan 2026.

View source on GitHub →

Wanich Risk Lab, 2026.