The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's primary workflow for creating presentations from scratch involves the agent generating multiple JavaScript files (e.g., slide-01.js, compile.js) and then executing them using Node.js. This runtime code generation and execution loop is a significant capability that could be misused if the generation process is influenced by malicious input.
[PROMPT_INJECTION]: The skill processes untrusted data from external PowerPoint files via the markitdown tool, which represents a surface for indirect prompt injection where instructions hidden in a presentation could influence subsequent agent actions.
Ingestion points: Content extraction from .pptx files via markitdown in SKILL.md and references/editing.md.
Boundary markers: No specific delimiters or safety instructions are used to distinguish ingested content from the agent's core instructions.
Capability inventory: The skill utilizes subprocess execution for Node.js and Python and performs extensive filesystem operations.
Sanitization: The documentation correctly recommends using defusedxml in references/editing.md for XML parsing to protect against XML External Entity (XXE) attacks.
[EXTERNAL_DOWNLOADS]: The skill requires several external dependencies including pptxgenjs, markitdown, and various Node.js packages. These are fetched from well-known repositories and registries.

pptx-generator