The Agent Skills Directory

[SAFE]: The skill provides structured instructions for generating content files within the readme-pages project. It restricts file creation to expected directories (src/content/notes/ and src/content/pages/) and does not request broad system access or execute arbitrary commands.
[PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it processes untrusted user input to create filesystem entries. Ingestion points: User input for 'Title', 'Slug', and 'Content' defined in SKILL.md. Boundary markers: No specific delimiters or safety warnings are included in the generated files to prevent content from being interpreted as instructions by an agent in the future. Capability inventory: Filesystem write operations for markdown files. Sanitization: The skill includes validation rules for titles, dates, tags, and descriptions, providing basic constraints on the ingested data.

create-content