The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to call local system utilities including soffice (for PDF conversion), pdftoppm (for image generation), and git (for document comparison). These operations are essential for the visual validation and text comparison features of the skill and utilize local file paths.
[EXTERNAL_DOWNLOADS]: The skill documentation lists requirements for several standard Python and Node.js packages, such as markitdown, pptxgenjs, and playwright, which are intended to be installed through standard package registries (npm/pip).
[SAFE]: The skill implements secure XML parsing by using the defusedxml library in its OOXML processing scripts (e.g., ooxml/scripts/base.py and ooxml/scripts/unpack.py), which effectively mitigates the risk of XML-based attacks like XXE.
[SAFE]: No patterns associated with prompt injection, credential exfiltration, or code obfuscation were identified during the analysis.
[SAFE]: The skill has an inherent attack surface for indirect prompt injection because it processes untrusted text content from user-provided PowerPoint files (Ingestion point: scripts/inventory.py; Boundary markers: absent; Capabilities: file system access and subprocess execution for image rendering; Sanitization: defusedxml for structure, none for natural language content). This is a standard risk for document-processing skills and is managed by core agent safety mechanisms.

pptx