A security researcher claims to have exploited a vulnerability in Perplexity AI, gaining access to what they describe as "unlimited Claude code." The researcher alleges that a single prompt and three shell commands were sufficient to compromise the system, suggesting this type of flaw may be present in other multi-agent AI systems. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Highlights potential security vulnerabilities in AI systems that could impact data privacy and system integrity.
RANK_REASON The article describes a claimed security exploit affecting a product, which falls under the 'tool' category.
