content-digest

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

• Third-party content exposure detected (high risk: 0.75). Runtime path: when the user provides a URL, the skill uses WebFetch to retrieve YouTube/article/podcast transcripts and then reads that fetched outsider-authored text into the agent for Stage 1 extraction and subsequent generation, creating an indirect prompt-injection exposure surface from arbitrary web content.