Hearmeman24

Build new blocks for BlockFlow, the visual pipeline editor for AI video and image generation. Use this skill whenever the user wants to create a new block, extend BlockFlow's functionality, add a pipeline step, build a custom node, or asks about the block architecture. Also trigger when the user mentions "custom_blocks", "frontend.block.tsx", "backend.block.py", block inputs/outputs, port types, or block registration. Even casual requests like "add a block that does X" or "I want a new pipeline step for Y" should use this skill.

Execute ComfyUI image/video generation workflows on remote RunPod servers using the comfy-gen CLI. Use this skill whenever the user wants to generate images or videos with ComfyUI remotely, submit workflows to a RunPod serverless endpoint, check generation status, download models, or configure remote ComfyUI settings. Also trigger when the user mentions ComfyUI workflows, RunPod image generation, remote GPU rendering, or serverless ComfyUI execution. Even if the user just says "generate an image", "run this workflow", or "submit to serverless", use this skill if a ComfyUI workflow is involved or implied. Trigger for both workflow execution and model management.

Deploy the HMAI LoRA Trainer serverless endpoint on RunPod and train LoRAs end-to-end on it — SDXL, Wan 2.2, Qwen-Image (+2512), FLUX.2 Klein 9B, Krea 2 (+Turbo), Z-Image Turbo, and Ideogram 4. Use this skill whenever the user wants to train a LoRA, fine-tune a diffusion model on their own images, deploy the aio-lora-trainer template, set up the HMAI trainer endpoint, or asks anything about dataset prep, config overrides, checkpoints, or retrieving trained .safetensors from this endpoint — even if they just say "train a LoRA of X" without naming the service.