google-dev-news-linebot

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

• Third-party content exposure detected (high risk: 0.95). The required workflow fetches RSS/Atom feeds from public sources at runtime (NEWS_SOURCES URLs), extracts their free-form title/summary text, and injects that outsider content into the Gemini prompt via build_prompt(...)->prompt and model.generate_content(prompt), creating an indirect prompt-injection path.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches RSS feeds at runtime from the listed URLs (e.g., https://developers.googleblog.com/feeds/posts/default) and directly injects the fetched article titles/summaries/links into the Gemini prompt used to generate ideas, so those remote feeds control the model input.