The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill documentation guides the agent to use the Slidev CLI for development and build tasks (e.g., slidev build, slidev export). It also details features like monaco-run which allows the execution of JavaScript and TypeScript code blocks directly within the presentation environment.- [DATA_EXFILTRATION] (LOW): The documentation describes the slidev --remote --tunnel command, which utilizes Cloudflare Quick Tunnels to make the local presentation server accessible via a public URL. This is a legitimate feature for remote sharing but requires user awareness of the network exposure.- [EXTERNAL_DOWNLOADS] (LOW): The skill identifies several external dependencies required for full functionality, including the playwright-chromium package for PDF rendering and various @iconify-json packages for icon support.- [PROMPT_INJECTION] (LOW): The skill introduces a surface for indirect prompt injection (Category 8) because it instructs the agent on how to generate slides from user input that can include executable code blocks or file-writing commands.
Ingestion points: Slide content generated from user prompts; external markdown files imported using the src attribute.
Boundary markers: Uses standard Markdown slide separators (---) and frontmatter blocks to delimit content.
Capability inventory: Browser-side code execution via monaco-run, file system writes via monaco-write, and network exposure via --tunnel.
Sanitization: No specific sanitization or filtering instructions are provided in the skill documentation; it relies on the default security model of the Slidev and Vite ecosystem.

slidev