The Agent Skills Directory

[SAFE]: The skill acts as a static repository of domain definitions (YAML-formatted) within Markdown files. It is intended to be read by other tools and does not perform any actions independently.
[PROMPT_INJECTION]: The expert_role definitions in domains/technical.md, domains/business.md, and domains/creative.md provide prompt templates for other agents. These templates use placeholders like {scope} and {context_summary} for string interpolation. While the skill itself is a passive data store, these templates constitute a surface for indirect prompt injection if the consuming agent populates them with untrusted data without implementing sanitization or boundary markers.

domain-registry