The Agent Skills Directory

[SAFE]: A comprehensive audit of all 15 files, including the primary skill definition and supporting reference implementations, confirms the skill is focused purely on architectural guidance and contains no malicious code or instructions.\n- [EXTERNAL_DOWNLOADS]: The skill templates reference legitimate, well-known Python packages for their intended purposes: sqlalchemy for database abstraction and uuid-utils for generating time-ordered UUIDv7 identifiers. No suspicious or unverified third-party dependencies are present.\n- [PROMPT_INJECTION]: The skill's instructions are strictly limited to domain modeling and architectural rules. There are no attempts to bypass safety filters, redefine agent identity in an unrestricted manner, or extract system prompts.\n- [DATA_EXFILTRATION]: No hardcoded credentials, API keys, or access to sensitive local file paths (such as .env or .ssh) were found. The skill does not perform any unauthorized data harvesting or outbound network transmissions of user data.\n- [REMOTE_CODE_EXECUTION]: The code provided consists of static templates intended for project scaffolding. There is no usage of dynamic execution sinks such as eval(), exec(), or runtime compilation from untrusted external sources.

domain-driven-design