The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use platform-specific shell commands (open, xdg-open, start) to open the generated HTML presentation in a web browser.
[EXTERNAL_DOWNLOADS]: For PowerPoint conversion, the skill attempts to use the python-pptx library and suggests installing it via pip if it is missing from the environment. This represents a standard dependency installation from a well-known repository.
[DYNAMIC_EXECUTION]: The skill generates and runs Python code to extract data from .ppt and .pptx files during the conversion process, which is a common pattern for file processing skills.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted data from external sources to generate executable HTML/JS code.
Ingestion points: User-provided notes, topics, and external PowerPoint (.ppt/.pptx) files processed in the conversion workflow.
Boundary markers: None identified in the instructions to delimit untrusted data from agent instructions during the content discovery or conversion phases.
Capability inventory: The agent can execute shell commands for file opening, run Python scripts for data extraction, and potentially use browser automation tools for layout validation.
Sanitization: No explicit sanitization, escaping, or validation steps are defined for the content extracted from external files or user notes before it is interpolated into the generated presentation code.

frontend-slides