JudgeMyAIPrompt Injection Red-Teaming: Curriculum Learning Tackles Cold-Start Problem Against Frontier Models
Curriculum reinforcement learning improves prompt injection red-teaming against frontier models like GPT-6-Luna and GPT-5.6-Terra.
BragJack: The Silent Hijacking of AI Browser Agents via Malicious Extensions — A Post-Mortem on the New Frontline of LLM Security
BragJack exploits reveal how malicious Chrome/Firefox extensions hijack AI browser agents (e.g., Perplexity, You.com, AI SearchGPT) via prompt injection. Deep dive into attack vectors, operational mitigations, and the $12M/year TCO of unpatched agent infrastructure.