The input file does not need to be strictly in a structurally valid PDF format. Instead, after reviewing the content, the original document disguised as a PDF will be sent to the parser. Malicious documents will construct malicious external entities that, through the protocol, point to local paths, thereby allowing access to any local files within the user's permission range.

The input file does not need to be strictly in a structurally valid PDF format. Instead, after reviewing the content, the original document disguised as a PDF will be sent to the parser. Malicious documents will construct malicious external entities that, through the protocol, point to local paths, thereby allowing access to any local files within the user's permission range.