The hackerbot-claw campaign. My take: AI is phenomenal in exploitation. Vulnerability research = craftsmanship, but exploitation? A guidebook trial-n-error loop, which agents excel at.
Maybe novel threat detection should flag LLM-generated content? 🤖