Brand

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly ingests and uses untrusted public references (e.g., SKILL.md Phase 2: "Manual references — User provides mood board images or links" and the tooling/fallbacks referencing public converters like freeconvert.com/png-to-svg and example URLs such as compost.fi/process.html), so external, user-provided web content is read and used to drive composition and decisions.