article-prompts-to-skills

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

• Third-party content exposure detected (high risk: 0.70). The required workflow includes “Read the repository instructions and the complete source article” at extraction time, meaning if that “source article” is outsider-authored content (e.g., pasted/fetched from the web or another party), its free-text prompts/criteria can be fed into the LLM context as part of the extraction and prompt writing.