LLM148 A LiteLLM 1.83.14 exploit chain leaks the master key through callback metadata, then abuses non-sandboxed Jinja2 GitLab prompts to achieve server-side RCE.
Claude AI91 The article shows how Claude Code plus MCP can automate vulnerability hunting with RE, fuzzing, RAG, bounty scoring, and strict validation gates…
Claude AI47 The article shows Claude turning a Phrack rsync exploit write-up into a faster working ARM64 RCE, rebuilding missing pieces, debugging without GDB,…
AI Security Research141 Researchers showed that the Claude AI model could generate a working exploit for a FreeBSD kernel vulnerability (CVE-2026-4747), producing a remote root…
