The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs platform-specific shell commands (open for macOS, xdg-open for Linux, and start for Windows) to display the generated HTML presentation to the user at the end of the workflow.
[EXTERNAL_DOWNLOADS]: For PowerPoint conversion functionality, the skill may request or perform the installation of the python-pptx library from the official Python Package Index (PyPI).
[EXTERNAL_DOWNLOADS]: Generated presentations may include references to external typography assets from established providers like Google Fonts and Fontshare.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted user-supplied notes and PowerPoint files to generate browser-executable code.
Ingestion points: Processes user-provided text drafts and uploaded .ppt or .pptx files during the conversion and creation phases.
Boundary markers: The skill does not explicitly define delimiters to separate untrusted user content from the generated presentation logic.
Capability inventory: Includes the ability to write files to the local system (HTML slides and assets), execute shell commands to open files, and potentially use browser automation for layout validation.
Sanitization: The instructions do not specify sanitization or escaping protocols for content extracted from PowerPoint files before it is injected into the HTML templates.

frontend-slides