Prompt Injection, Esq.: pwning AI agent guardrails with lawyerspeak
Comments
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Summary
A security engineer at Opsin Security recounts how a read-only ChatGPT workspace agent that manages their Notion hiring tracker was talked into editing records after a fake law-firm letter, generated with the Heavyweight tool, was posted in Slack. The agent followed the official-looking letter over its original guardrail.