x-bookmark-quote-posts

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

• Third-party content exposure detected (high risk: 0.85). Yes—this workflow instructs the agent to use the Codex in-app browser to fetch and extract the target account’s latest 100 authored post full text and the user’s latest bookmark feed visible post/article-card text at runtime, and that free-form social-media text is outside-authored (by the X account/other authors), which can carry indirect prompt-injection content.