The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to invoke system binaries including soffice (LibreOffice) and pdftoppm. These calls are implemented using list arguments rather than shell strings, which effectively mitigates the risk of command injection. The usage is strictly limited to the skill's primary purpose of rendering slide previews and final decks.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from .pptx templates and user-provided speaker notes. It implements robust defenses, including a specific instruction for the agent to treat this content strictly as data and ignore any instructions found within it. Additionally, it uses structured XML parsing (xml.etree.ElementTree) and filters illegal XML control characters, reducing the risk of data-driven attacks.
[DATA_EXPOSURE]: The fit.py script scans system font directories (e.g., /Library/Fonts, C:\Windows\Fonts) and user-specified paths in the PPTX_STYLE_FONT_DIRS environment variable. This is a legitimate functional requirement to ensure accurate text-fit measurements using real font metrics.
[SAFE]: The skill is well-structured and follows best practices for document manipulation. It focuses on XML surgery within the OpenXML package format and does not perform network operations or access sensitive system configuration files.

pptx-style-build