Tag: Prompt Injection

  1. A honeypot site tricked Claude's web_fetch into leaking user data (simonwillison.net)
    Security · · July 16, 2026
  2. 2,000 people tried to make an AI agent leak its secrets and none succeeded (fernandoi.cl)
    AI · · June 27, 2026
  3. Prompt injection works because models judge text by how it sounds (arxiv.org)
    AI · · June 23, 2026
  4. A 2-cent transfer reveals prompt injection in Bunq's banking AI (blue41.com)
    AI · · June 11, 2026
  5. OpenAI ships a Lockdown Mode against prompt injection exfiltration (simonwillison.net)
    Security · · June 7, 2026
  6. Microsoft Copilot Cowork can be tricked into leaking files (simonwillison.net)
    AI · · May 26, 2026

← all tags