source-summarization

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

• Third-party content exposure detected (high risk: 0.70). The skill ingests meta.md and transcript.md/content.md from a user-provided source folder; if those files are authored by outsiders (e.g., downloaded/forwarded documents or transcripts), their free-form text is read at runtime and becomes LLM context during the section-by-section summarization.