stack-guide

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). This skill explicitly exposes the agent to arbitrary web content via the gstack headless/browser tools (e.g., /gstack-browse, /gstack-connect-chrome and the /gstack-qa and /gstack-design-review workflows that navigate, screenshot, and interact with pages), which means untrusted third-party pages could be read and influence agent decisions.