AI Agent Introduced A Flaw in Snowflake’s Code—Then Another AI Agent Exploited It

An artificial intelligence-generated security fix introduced a vulnerability into one of Snowflake’s public GitHub repositories, before a separate autonomous AI security agent found the defect, exploited it and extracted credentials providing access to parts of the company’s internal Jira environment.

The incident did not result from an uncontrolled attack or criminal operation. It occurred during authorised security research conducted by Wiz through Snowflake’s HackerOne vulnerability disclosure programme.