document-parser

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md usage examples explicitly show accepting a remote URL (e.g., document-parser "https://example.com/doc.pdf") and the skill's purpose is to extract and return document content (extract.py extract(path) returns extracted text/metadata), so it clearly ingests untrusted public web documents that could contain instructions influencing agent behavior.