The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands such as python, pdftoppm, and soffice to extract text, unpack XML, and convert files to images. These operations involve processing user-supplied .pptx files.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its document processing workflows.
Ingestion points: External .pptx files are ingested and converted to text via markitdown.
Boundary markers: No markers or instructions are provided to the agent to distinguish between presentation content and operational commands.
Capability inventory: The skill possesses capabilities to execute Python scripts, system binaries like LibreOffice and Poppler, and file system operations.
Sanitization: There is no evidence of sanitization or filtering of the content extracted from presentations before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill depends on standard external packages including markitdown, Pillow, and pptxgenjs from public registries.
[NO_CODE]: Several internal scripts (scripts/thumbnail.py, scripts/office/unpack.py, scripts/office/soffice.py) and referenced markdown guides (editing.md, pptxgenjs.md) are mentioned in the skill instructions but are missing from the file package, preventing a full code audit.

pptx