A Python framework for combining static analysis with LLM-based bug verification
-
Updated
Oct 4, 2026 - Python
A Python framework for combining static analysis with LLM-based bug verification
AI 驱动的漏洞自动验证平台:提交仓库与漏洞描述,在隔离沙箱中白盒审计、搭靶场复现并出具中文报告。
This project, developed with Python and PyQt6, involves creating an .exe application to verify and change vulnerability statuses, visualize results, and export new vulnerabilities.
Stage338: Behavior Decision Engine for expected-vs-actual behavior verification.
REMEDA Stage327: structured reproduction evidence schema for AI vulnerability verification, target matching, SHA256 integrity binding, and third-party audit verification.
AI vulnerability verification demo using the Stage299 QSP/VEP Gate engine.
agy-security-audit is an evidence-backed, multi-stage security assurance and vulnerability verification plugin for Antigravity CLI (agy).
Safe Reproduction Template Library for AI vulnerability verification. Defines safe reproduction templates, expected behavior, pass/fail conditions, and safety boundaries without attack code.
To associate your repository with the vulnerability-verification topic, visit your repo's landing page and select "manage topics."