super-extract

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill mandates emitting exact verbatim substrings from source documents (extraction_text in JSON and highlighted HTML), so if a document contains secrets (API keys, tokens, passwords) the LLM will be forced to output them verbatim, creating a high exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly downloads and ingests documents from arbitrary URLs (see "Step 0: If the source is a URL to a PDF, first download it" and the use of pdf/docx skills to extract text), meaning untrusted third‑party content is read and interpreted as part of the extraction workflow and can change extraction/processing decisions.