bouncer-feed-filter

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

• Third-party content exposure detected (high risk: 0.90). The extension's content script scrapes user-generated posts from Twitter/X (src/content/feedObserver.ts and src/adapters/twitter.ts) and sends the post text/images into the AI classification prompt (src/models/classify.ts / buildClassificationPrompt), allowing untrusted third‑party content to directly influence model decisions and subsequent actions (hiding posts) in the workflow.