paw-mkt-social

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to audit live public URLs and run live competitor research using agent-browser (see "Starting Context Router" in references/shared-patterns.md and the note in references/analytics.md), which requires fetching and interpreting untrusted public web/social content (e.g., competitor sites, social platforms, Reddit) and that content is used as the working source of truth to drive decisions and actions.