The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references standard Astro CLI commands for project creation (npm create astro@latest) and integration management (npx astro add). These utilize official, well-known registries and are recognized as safe ecosystem practices.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as Markdown/MDX files and user-provided API request bodies, creating a potential injection surface.
Ingestion points: The src/content/ directory for content collections and Astro.request.json() in SSR API routes.
Boundary markers: The documentation recommends using import.meta.env.PUBLIC_* to clearly delineate environment variables intended for the client versus the server.
Capability inventory: Standard Astro framework capabilities including client-side fetching and server-side execution of data-fetching logic.
Sanitization: The skill explicitly warns users about the security risks of the set:html directive and provides instructions on validating inputs in SSR mode to mitigate common vulnerabilities.
[SAFE]: No malicious patterns, obfuscation techniques, hardcoded credentials, or unauthorized persistence mechanisms were identified. The instructions follow established web development safety standards.

astro