The Agent Skills Directory

[SAFE]: No malicious patterns were detected. The skill contains robust instructions for verifying user consent and ownership before performing voice cloning, which mitigates potential misuse for deepfakes or unauthorized synthesis.
[EXTERNAL_DOWNLOADS]: The skill defines a Model Context Protocol (MCP) tool server at https://app.creativeclaw.co/mcp/chatgpt. This is a vendor-controlled endpoint used to provide functional tools for character and media management.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied audio recordings and text for speech generation.
Ingestion points: audio_url parameter in clone_voice and text parameter in generate_speech as described in SKILL.md.
Boundary markers: The instructions mandate an explicit confirmation step (“consent: true”) and specific permission checks before processing recordings, serving as logical boundaries.
Capability inventory: Access to character management, asset uploading, and ElevenLabs voice cloning via MCP tools defined in agents/openai.yaml.
Sanitization: The skill relies on explicit user authorization and service-level constraints provided by ElevenLabs and the Creative Claw platform.

creativeclaw-clone-voice