daily-news-digests

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s required workflow reads outsider-authored free text by collecting and ingesting items from public sources it monitors at runtime (e.g., GitHub issues/PRs and Reddit public search feeds), so an outsider can submit poison content into those sources for the agent to fetch and summarize.