The Agent Skills Directory

[SAFE]: The skill defines a coordination and routing framework for a Direct-to-Consumer (DTC) business system. It does not contain shell commands, script execution, or network exfiltration logic.
[DATA_EXPOSURE]: The skill identifies internal project files such as products.md and brand-brain as data sources for business audits. These are domain-specific configuration files and do not include system-level sensitive information like SSH keys or API tokens.
[INDIRECT_PROMPT_INJECTION]: The skill processes user_request data from a parent 'Hub' module. While it lacks explicit sanitization instructions for this input, its capabilities are limited to selecting internal routing paths (afa-ops, afa-expand) and generating structured YAML reports. No dangerous primitives (e.g., code execution or external network calls) are exposed to the processed data.

afa-scale