dingtalk-workflow-social-tracker

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

• Third-party content exposure detected (high risk: 0.80). The skill's Mode 3 "自动数据采集" workflow (Step 2) explicitly performs AI WebSearch against public sites (e.g., "WebSearch → Social Blade 数据") and parses those open/public search results to update snapshots and trigger notifications/decisions, meaning untrusted third‑party content is ingested and can influence agent actions.