brief-parser

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly accepts URLs/links and, in its required Processing Pipeline Step 0.5, instructs searching and extracting content from public social platforms (小红书/抖音/B站/微博 and mentions X/Reddit/etc.), so it ingests untrusted, user-generated third‑party content that will influence analysis and creative decisions.