The Agent Skills Directory

[SAFE]: The skill operates within a defined project structure, reading creative inputs from subdirectories like project/screenplay/ and project/characters/, and writing generated outputs to project/prompts/. This is standard behavior for an AI agent integrated into a creative workflow.
[SAFE]: No evidence of prompt injection or attempts to bypass AI safety filters was found. The instructions emphasize preserving upstream creative decisions and providing "safe alternatives" for risky scenes.
[SAFE]: No obfuscation techniques, such as Base64 encoding of commands or hidden Unicode characters, are present in the skill files.
[SAFE]: The skill references well-known and legitimate AI services (e.g., Midjourney, DALL-E, Sora, Runway) for optimization purposes and does not attempt to download or execute external scripts or packages.
[SAFE]: No hardcoded credentials or unauthorized data exfiltration patterns were detected. Network operations are limited to the expected use of the agent's provided tools.

creator-prompt-engineer