The Agent Skills Directory

[SAFE]: No security threats were identified. The skill instructions focus entirely on copywriting quality and editorial standards, utilizing structured workflows and rubrics for evaluation.
[SAFE]: The skill includes a functional 'Learning Loop' that ingests human-edited versions to improve future copy. While this represents a data ingestion surface, the process is internal to the skill's logic and includes safeguards such as manual transfer testing and durable learning gates to maintain strategic integrity.
[SAFE]: No network exfiltration, remote code execution, or prompt injection attempts were found. All referenced dependencies and tools are internal to the agent platform's skill ecosystem.

compound-copywriting