https://github.com/uiuc-kang-lab/cve-bench

CVE-Bench: A Benchmark for AI Agents’ Ability to Exploit Real-World Web Application Vulnerabilities
https://github.com/uiuc-kang-lab/cve-bench

benchmark inspect language-model

Last synced: 4 months ago
JSON representation

CVE-Bench: A Benchmark for AI Agents’ Ability to Exploit Real-World Web Application Vulnerabilities

• Host: GitHub
• URL: https://github.com/uiuc-kang-lab/cve-bench
• Owner: uiuc-kang-lab
• License: apache-2.0
• Created: 2025-02-25T16:35:57.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2026-01-10T05:06:06.000Z (5 months ago)
• Last Synced: 2026-01-11T01:32:53.134Z (5 months ago)
• Topics: benchmark, inspect, language-model
• Language: Python
• Homepage: https://arxiv.org/abs/2503.17332
• Size: 121 MB
• Stars: 136
• Watchers: 3
• Forks: 29
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-offensive-security-ai - CVE-Bench GitHub
• Awesome-AI-Security - CVE-Bench - kang-lab/cve-bench?logo=github&label=&style=social)](https://github.com/uiuc-kang-lab/cve-bench) - 40 dockerized web CVEs; success = expected impact triggered. [arXiv](https://arxiv.org/abs/2503.17332) ([↑](#table-of-contents)Benchmarks <a name="benchmarking"></a> / **Adversarial Resilience**)
• awesome-rainmana - uiuc-kang-lab/cve-bench - CVE-Bench: A Benchmark for AI Agents’ Ability to Exploit Real-World Web Application Vulnerabilities (Python)