last30days

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests public, user-generated content from Reddit, X, YouTube, TikTok, Hacker News, Polymarket, GitHub and the web as part of its mandatory research workflow (see the SKILL.md description and required steps Step 0.55 / Step 1 / Step 2), and that external content is read and synthesized to drive recommendations, prompts, and next actions—allowing indirect prompt injection via third‑party text.