The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from PowerPoint files (.ppt, .pptx) and user-supplied notes to generate code-heavy HTML presentations.
Ingestion points: Content discovery and PPT conversion workflows in SKILL.md.
Boundary markers: Absent. There are no instructions to use delimiters or ignore embedded instructions within the processed content.
Capability inventory: The agent can execute shell commands (python3, open), write files (presentation.html), and create directories (.ecc-design/slide-previews/) across SKILL.md and related processes.
Sanitization: Absent. Extracted text and images are used directly in code generation without validation or escaping.
[COMMAND_EXECUTION]: The skill utilizes system commands to provide its primary features.
Executes python3 to run extraction scripts for PowerPoint files.
Uses platform-native commands (open, xdg-open, start) to launch generated files in the user's browser.
[EXTERNAL_DOWNLOADS]: The skill links to well-known external design services.
References Google Fonts and Fontshare for typography in generated presentations.
[DYNAMIC_EXECUTION]: The skill dynamically generates and manages executable web content.
Assembles HTML, CSS, and JavaScript presentations at runtime based on user templates and content.
Creates and manages temporary preview files in the .ecc-design/ directory.

frontend-slides