More dynamic parameters for ComfyUI Txt2Img

[yt-koike]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

1. Is this request related to a challenge you're experiencing? Tell me about your story.

Hello, this is my first discussion in the Dify project.
I have been using this wonderful software to implement an AI agent on Slack.
Especially ComfyUI is a very good tool to generate images but I found there are many static parameters in the ComfyUI nodes, which cannot be modified by LLM's results or so.

For example, txt2img node has some static parameters such as inference steps, image resolution(width&height) and negative prompts. I think these can be easily converted to dynamic ones as they are just integers or strings.

Meanwhile, LORAs, sampling methods and schedular could be a bit more tricky as they need a specific name, not an arbitrary string.

2. Additional context or comments

No response

3. Can you help us with this feature?

• I am interested in contributing to this feature.

[yt-koike]

I think we can implement this by changing form: form into form:llm in comfyui_stable_diffusion.yaml