
Abstract On July 29th, 2026 a study was performed to see if LLM generated Opengrep rules could outperform Semgrep’s community rules when ran against OWASP’s Python SAST Benchmark program. To limit scope, patterns for file path traversal vulnerabilities were only focused on. Gemini 3.1 Pro (Low Thinking Mode) was used to generate static analysis rules […]
One Up Security, LLC


