Wiz AI Agent Exploits Snowflake GitHub Vulnerability

An autonomous AI security agent developed by cloud security firm Wiz identified and exploited a critical vulnerability in a Snowflake GitHub Actions workflow, ultimately gaining access to the company's internal Jira credentials. The incident, which occurred in June 2026, highlights the dual-edged nature of AI in software development, where one AI tool can introduce a flaw while another can autonomously discover and exploit it.

The vulnerability was discovered by Wiz's Red Agent, an autonomous security research tool, while scanning Snowflake's public GitHub organization as part of a HackerOne disclosure program. The flaw resided in the jira_issue.yml workflow of the snowflakeconnectornet repository. This workflow was designed to run whenever a GitHub issue was opened, using the issue's title as part of a shell command. A code change introduced in a pull request on June 18, 2026, replaced a safer input-handling pattern with a direct string expansion, creating a script injection vulnerability. A guard condition intended to prevent exploitation was ineffective because it was built for pull request events, not issue events, allowing any GitHub user to bypass it.