The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes a Model Context Protocol (MCP) tool named creative-claw at https://app.creativeclaw.co/mcp/chatgpt. This is a vendor-owned resource hosted on the official creativeclaw.co domain.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external media assets via URLs, which creates an ingestion point for untrusted data. While this presents a potential surface for indirect prompt injection, the skill defines a rigid prompt structure to isolate these references.
Ingestion points: The workflow in SKILL.md ingests external data through the image_url, last_frame_url, image_urls, video_urls, and audio_urls parameters.
Boundary markers: The skill enforces a structured prompt template (References, Shot, Subject, Action, Camera, Look, Audio, Continuity) which acts as a boundary to separate asset citations from instructions.
Capability inventory: The skill calls get_model_params and submit_feedback via the creative-claw MCP tool to perform video generation tasks.
Sanitization: There is no evidence of sanitization or validation of the content contained within the referenced media assets.

creativeclaw-minimax-h3-max