Skill: comfyui-prompt-automation-orchestrator

Description

Automates the end-to-end pipeline of extracting prompts from Markdown files, injecting them into ComfyUI 'Save' format JSON workflows, and executing them via the Direct API.

Trigger Conditions

• When the user asks to "run the [Meal Name] prompts" or "generate images/video for [Meal Name]".
• When a batch of content needs to be pushed to the remote ComfyUI server.

Workflow

1. Locate Files: Identify the .md prompt file and the target .json workflow in the Food Prep Content directory.
2. Parse: Use prompt_parser_util.py to extract text blocks from the Markdown.
3. Inject: Use the parser to inject the text into the target node's widgets_values.
4. Wrap & Send: Convert the "Save" format JSON into the "prompt": { ... } API format and use curl to POST to the remote ComfyUI endpoint.

Pitfalls

• Node IDs: Ensure the Python script targets the correct id or type within the JSON.
• JSON Wrapping: Failure to wrap the 'Save' format into a 'prompt' object will result in 400 errors from ComfyUI.
• VRAM: High-resolution video tasks (Wan 2.2) require careful monitoring of the remote server's status.