The Agent Skills Directory

[SAFE] (SAFE): The skill is a standard utility for document processing and validation. It uses standard Python libraries and follows defensive coding practices, specifically using the defusedxml library for all XML parsing to protect against XML External Entity (XXE) attacks.
[Indirect Prompt Injection] (LOW): The skill processes untrusted Office documents. A vulnerability surface exists in ooxml/scripts/unpack.py and ooxml/scripts/validation/docx.py where zipfile.extractall() is used without path validation. This creates a 'Zip Slip' (path traversal) risk if a maliciously crafted Office document is processed by the agent. Additionally, ooxml/scripts/pack.py executes the soffice command for document validation.
Ingestion points: ooxml/scripts/unpack.py and ooxml/scripts/validation/docx.py (via zip file extraction).
Boundary markers: Absent.
Capability inventory: File system write access via extraction and local command execution via soffice.
Sanitization: defusedxml is used for XML content, but ZIP member paths are not sanitized.

pptx