blog-content-enhancer

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

• Third-party content exposure detected (high risk: 0.90). The skill's research workflow explicitly performs WebSearch and compiles citations from open/public sites (see the research command and references/research-sources.md which list Reddit, Hacker News, TechCrunch, arXiv, etc.), so the agent will fetch and read untrusted, user-generated third-party content as part of its routine.