og-image-design

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow shown in SKILL.md calls belt app run infsh/html-to-image with user-supplied JSON containing an html string (free text) that the service would render to an image, so outsiders can provide arbitrary HTML text directly into the monitored input field.