thumbnail-creator

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

• Third-party content exposure detected (high risk: 0.85). The skill ingests “Article copy or URL” from the operating user and, if a URL is provided, fetches that page at runtime to extract paragraphs/H1; this is outsider-authored free text from arbitrary public web content that the LLM reads into context for composition/prompting.