The Agent Skills Directory

[SAFE]: The skill documents legitimate framework features, architectural patterns, and development workflows for the Astro ecosystem without introducing malicious behavior.
[INDIRECT_PROMPT_INJECTION]: The skill identifies surfaces for ingesting external data, such as fetching from third-party APIs in references/ssg-patterns.md and loading content from external sources in references/content-collections.md. These patterns represent a standard web development attack surface. The skill proactively addresses these risks by providing patterns for strict schema validation with Zod in references/actions.md and references/env-variables.md, and by detailing secure environment variable management to prevent accidental data exposure.

astro