@doany-ai

video-extend chains or extends video clips using google veo 3-1 endpoints via the runcomfy cli, supporting narrative shot-by-shot generation from a seed prompt.

ai-avatar-video routes text and audio to multiple runcomfy avatar models (omnihuman, wan, happyhorse, seedance) with lip-sync, returning invoke commands and prompts for the selected path.

image-outpainting extends images beyond original boundaries using runcomfy endpoints, supporting prose-driven extension, reference matching, and brand-locked continuation.

elevenlabs-music-generation wraps the runcomfy cli to produce multi-minute audio tracks from text prompts, but lacks actionable procedure, error handling, and concrete input/output contracts.

flux-kontext wraps black forest labs flux 1 kontext pro via runcomfy for single-image local edits. lacks procedural depth, executable steps, and failure mode documentation.

nano-banana-2 wraps google's flash-tier text-to-image model via runcomfy, optimized for rapid ideation and social thumbnails. lacks procedural substance and actionable step-by-step guidance.

>

>

>

>

>

runcomfy-cli — an installable skill for AI agents, published by doany-ai/skills.

>

relight — an installable skill for AI agents, published by doany-ai/skills.

>

video-inpainting — an installable skill for AI agents, published by doany-ai/skills.

>

>

wan-2-7 — an installable skill for AI agents, published by doany-ai/skills.

>

>

>

>

video-outpainting — an installable skill for AI agents, published by doany-ai/skills.

>

>

>

>

>