The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands for environment configuration and file processing.
Installs packages using npm and pip during the setup phase.
Uses pdftoppm for PDF-to-image conversion and a system-provided soffice.py script for document transformation.
[EXTERNAL_DOWNLOADS]: The skill downloads several well-known and standard libraries from official package registries.
Fetches Python package markitdown for document extraction.
Fetches Node.js packages pptxgenjs, react, react-dom, react-icons, and sharp for slide generation and image processing.
[REMOTE_CODE_EXECUTION]: The skill generates and executes a custom JavaScript file (generate.js) to handle the presentation building logic.
The script uses the pptxgenjs library to programmatically create the slide deck file.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the ingestion of untrusted external content.
Ingestion points: Text content is extracted from user-supplied PDF and Word documents using the markitdown utility.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when processing extracted text.
Capability inventory: The environment allows file system writes, shell command execution for quality assurance, and dynamic execution of generated JavaScript code.
Sanitization: There is no explicit requirement to sanitize or escape the content extracted from external documents before it is used in the slide generation logic.

slide-deck-builder