Guardian is an enterprise-level AI penetration testing tool by Zakir Kun on GitHub. It combines OpenAI GPT-4, Anthropic Claude, Google Gemini, and OpenRouter in a single multi-agent system for smart security assessments and thorough evidence collection. It employs four active agents: Planner, Tool Selector, Analyst, and Reporter. They collaborate on …

OpenAI along with crypto firm Paradigm have launched EVMbench, a benchmark to assess AI agents’ skills in identifying, fixing, and exploiting serious vulnerabilities in smart contracts. EVMbench includes 120 curated vulnerabilities from 40 security audits, primarily from open code audit competitions on platforms like Code4rena. Three Evaluation Modes: EVMbench tests …

A China-linked cyberespionage group has exploited a zero-day vulnerability in Dell’s RecoverPoint for Virtual Machines since at least mid-2024, according to Google’s Threat Intelligence Group and Mandiant. GTIG and Mandiant attributed the exploitation of CVE-2026-22769 to a group called UNC6201, who used the vulnerability for lateral movement, persistence, and deploying …