The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill and its generated templates reference and fetch resources from well-known and trusted services.
The build configuration in assets/template/vite.config.ts performs a fetch request to the official NPM registry (registry.npmjs.org) to retrieve package version metadata.
The landing page template in assets/template/index.html includes preconnect and stylesheet links for Google Fonts (fonts.googleapis.com and fonts.gstatic.com).
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied product information—such as the product name, description, and feature lists—and interpolates these strings into the generated React components. This pattern creates a surface for indirect prompt injection if malicious instructions are embedded in the user-provided product data.
Ingestion points: User-provided metadata used for string replacements as instructed in SKILL.md across various template files in assets/template/src/components/sections/.
Boundary markers: There are no explicit delimiters or instructions defined in the template files to guard against instructions embedded within the interpolated strings.
Capability inventory: The skill automates the creation of a Vite project and populates it with source code, but it does not execute the resulting code in the host environment.
Sanitization: The skill relies on the agent's interpretation and standard React JSX escaping; it does not implement custom sanitization for the interpolated data.

tech-product-landing