docx

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly unpacks and parses arbitrary .docx files provided by users—see ooxml/scripts/unpack.py, the pandoc text-extraction commands in SKILL.md, and the Document/document.xml processing in the redlining workflows—so the agent will read and interpret untrusted user-supplied document content that could contain injected instructions.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). The prompt explicitly instructs installing system packages (e.g., "sudo apt-get install pandoc", "sudo apt-get install libreoffice", global npm installs) and other system-level dependency changes that require elevated privileges and can modify the machine state, even though it doesn't instruct creating users or bypassing security mechanisms.