The Agent Skills Directory

[SAFE]: The skill serves as a meta-tool providing guidelines and documentation for skill development. It defines structures, frontmatter requirements, and writing principles for other developers.
[DYNAMIC_CONTEXT_INJECTION]: The skill references and documents the !command syntax (dynamic context injection) used in platforms like Claude Code. These references are purely educational, providing examples like git diff for development workflows. Per the security analysis guidelines, documentation and tutorials for this syntax are not considered suspicious.
[INDIRECT_PROMPT_INJECTION]: The skill ingests a user-provided domain via the $ARGUMENTS variable to generate a skill template. This is a standard text-transformation task with a clearly defined instruction set and does not represent a significant vulnerability surface.
[METADATA_POISONING]: The frontmatter fields (name, description) are used correctly to identify the skill's purpose and activation triggers, following the recommended standard without deceptive intent.

skill-creator