Tag
1 article
Security researcher Aonan Guan exploited prompt injection flaws in AI agents from Anthropic, Google, and Microsoft, stealing API keys. All three companies paid bug bounties but did not issue public advisories.