afa-pr

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflows (e.g., "声誉监控与危机响应", "品牌资产保护" and diagnostic/monitoring steps) explicitly instruct ingesting and acting on open/public third‑party content — naming platforms like Twitter/X, Reddit, Trustpilot, Google News, Amazon, TikTok and requiring collection of URLs/screenshots/WHOIS lookups — and uses that external content to determine crisis level and drive takedowns and response actions.