competitor-watch

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md: Step 2 uses browser_json(url, prompt=...) and browser_scrape(url, ...) to fetch and extract readable content from user-supplied competitor URLs at runtime, which is outsider-authored free text (public web pages) that can be fed into the LLM for extraction.