The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of structured data from both local and remote sources, which is the primary purpose of the Astro Content Layer.
Ingestion points: Data is retrieved from local directories via glob(), local files via file(), and external APIs through custom loaders in src/content.config.ts.
Boundary markers: The skill mandates the implementation of Zod schemas to define data structures, which acts as a validation boundary for incoming content.
Capability inventory: The implementation involves filesystem read access for local files and network capabilities via fetch for remote data retrieval.
Sanitization: Zod 4 validation is an enforced best practice in the skill guidelines to ensure entry data integrity.
[EXTERNAL_DOWNLOADS]: The skill references standard mechanisms for fetching remote data and installing dependencies.
The custom loader template demonstrates fetching content from a remote API endpoint using fetch().
The documentation includes standard setup procedures for official integrations using the npx package runner.
[COMMAND_EXECUTION]: The skill documentation includes standard CLI usage for environment setup.
Mentions the use of npx astro add mdx to add official Astro integrations.
Refers to the astro sync command for generating project-specific TypeScript types.

astro-content