xsearch

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored free text from X (tweet author text/handles/URLs) is extracted by browser-harness-js via Runtime.evaluate on https://x.com/search?... and then returned to the skill output, which is then printed/returned to the agent (i.e., LLM context can include that scraped tweet text).