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 user-generated content from public sites (Reddit and X) and general web search as part of its required workflow (see SKILL.md "Research ANY topic across Reddit, X, and the web" and SPEC.md / scripts/last30days.py which call openai_reddit.search_reddit, xai_x.search_x and a WebSearch tool), and then the agent is instructed to read, synthesize, and let those findings determine prompt formats and follow-up actions—so untrusted third‑party content can materially influence tool behavior.