The Agent Skills Directory

[COMMAND_EXECUTION]: Internal orchestration scripts such as generate.py, edit.py, validate.py, and profile.py utilize subprocess.run to execute specialized Python modules within the same skill package. These calls are constructed using the current interpreter and absolute path resolution, ensuring controlled execution of internal logic.
[COMMAND_EXECUTION]: The slide rendering and validation components (thumbnail.py, visual_validator.py) execute local system utilities including soffice (LibreOffice) and pdftoppm (Poppler) to handle PowerPoint-to-PDF conversion and image generation. These are essential for the skill's primary functionality of visual validation.
[COMMAND_EXECUTION]: The font_fallback.py utility uses subprocess.run to call fc-match for discovering available system fonts to ensure consistent text measurement across platforms.
[PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection as it ingests and processes untrusted markdown content and JSON replacement data.
Ingestion points: Primary input is through ingest.py (parsing markdown outlines) and edit.py (processing JSON text replacements).
Boundary markers: The parser relies on structural markdown elements such as slide headers (# Slide N:) and separators (---) to delimit slide boundaries and content types.
Capability inventory: The skill possesses the ability to write to the file system (prs.save) and execute local binaries via subprocess.run for rendering tasks.
Sanitization: Input content is cleaned and validated using a state-machine parser and regular expressions to extract text runs and typography markers, ensuring that inputs are treated as data rather than executable instructions.

pptx-from-layouts