quant-research-replication

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow ingests outsider-authored free text when it performs source discovery and then downloads/opens PDFs/webpage text provided or found from user-supplied search topics (via scripts/source_discovery.py fetching web pages/snippets and potentially downloading PDFs), before generating the translation and factor reconstruction outputs.