The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill manages content within Astro MDX collections, which introduces an ingestion surface for untrusted data if blog posts are sourced from external contributors.
Ingestion points: MDX content files and collection schemas defined in SKILL.md and references/technical.md.
Boundary markers: The skill recommends using Zod schemas for data validation and centralized publication filters to distinguish between drafts and public content.
Capability inventory: File system read access for project configuration and execution of the astrofy verification tool.
Sanitization: Implementation of publication status checks (drafts vs. published) and schema validation via Zod loader in src/content.config.ts.
[COMMAND_EXECUTION]: The skill utilizes a command-line utility astrofy to perform category-specific project checks.
Evidence: The instruction to run astrofy check --category blog appears in SKILL.md and examples/cases.md. This tool is part of the developer workflow for validating the blog structure and publication rules within the local environment.

astrofy-blog