The Agent Skills Directory

[SAFE]: The skill instructions were thoroughly reviewed across all threat categories including prompt injection, data exfiltration, and remote code execution. No malicious behaviors, obfuscation techniques, or unauthorized resource accesses were identified. The workflow includes a mandatory confirmation step (Phase 2), which enforces human-in-the-loop verification before the AI proceeds with content generation. The skill does not utilize any external dependencies, shell commands, or network operations.

copywriting