higgsfield-image-auto

这段内容是审核时保存的快照。外部来源才是完整且最新的版本。

---
name: higgsfield-image-auto
description: Automatically generate an AI image on Higgsfield using Playwright browser automation. Use when the user has an image prompt and wants to generate it on Higgsfield Soul 2.0 or Nano Banana Pro. Triggers on requests like "generate image on higgsfield", "create image", "auto-generate image", "make the image on higgsfield", or any request to submit an image generation job. Requires Playwright MCP tools.
---

# Higgsfield Image Auto-Generate via Playwright

This skill automates the full Higgsfield image generation flow using Playwright browser automation. It navigates to the image generation page, enters a prompt, configures settings, and clicks Generate — hands-free.

---

## Prerequisites

- User must be **logged in** to higgsfield.ai in the Playwright browser session
- **Playwright MCP** plugin must be enabled
- Load each Playwright tool with `ToolSearch` before first use

## Input Requirements

The user provides:
1. **A text prompt** for the image (or invoke `/ugc-hot-girl` first to generate one)

Optional overrides (defaults shown):
- **Model**: `Soul 2.0` (options: Soul 2.0, Soul Cinema, Nano Banana Pro, Nano Banana 2, Seedream 5.0 lite, GPT Image 1.5, Grok Imagine, FLUX.2, Reve, Z-Image)
- **Aspect Ratio**: `3:4` (options: 1:1, 3:4, 4:3, 9:16, 16:9, 2:3, 3:2)
- **Resolution**: `2k` (options: 1k, 2k, 4k)
- **Image Count**: `1` (options: 1-4)

---

## Model URL Map

Each model has its own URL path:

| Model | URL |
|---|---|
| Soul 2.0 | `/image/soul-v2` |
| Soul Cinema | `/image/soul-cinematic` |
| Nano Banana Pro | `/image/nano-banana-pro` |
| Nano Banana 2 | `/image/nano-banana-2` |
| Seedream 5.0 lite | `/image/seedream_v5_lite` |
| GPT Image 1.5 | `/image/openai_hazel` |
| Grok Imagine | `/image/grok_image` |
| FLUX.2 | `/image/flux_2` |
| Reve | `/image/reve` |
| Z-Image | `/image/z-image` |

**Default**: Soul 2.0 (`/image/soul-v2`) — best for photorealistic portraits and UGC characters.

---

## Automation Flow

### Step 1: Navigate to the image generation page

```
browser_navigate → url: "https://higgsfield.ai/image/soul-v2"
```

Replace `soul-v2` with the appropriate model path from the table above if the user requests a different model.

Wait for the page to load, then take a snapshot to confirm:
- The page title contains the model name
- The prompt textbox is visible

### Step 2: Find and click the prompt textbox

Take a snapshot. Look for:
```
textbox [ref=eXX]
```
It's inside `group [ref=eXX]` at the bottom of the page. The placeholder text is "Describe the scene you imagine".

The textbox has a Playwright ID: `[id="hf:tour-image-prompt"]` — Playwright may use this selector automatically.

```
browser_click → ref: <textbox_ref>, element: "Image prompt textbox"
```

### Step 3: Type the prompt

```
browser_type → ref: <textbox_ref>, text: "<the image prompt>"
```

**Important**: Use the `fill` method (default), NOT `slowly: true`. The textbox is a standard input, not a Lexical editor.

### Step 4: Adjust settings (optional)

If the user wants non-default settings, click the corresponding buttons:

**Aspect Ratio** — Find the button showing current ratio (e.g., "3:4"):
```
browser_click → ref: <ratio_button_ref>, element: "Aspect ratio selector"
```
Then select from the dropdown options.

**Resolution** — Find the button showing current resolution (e.g., "2k"):
```
browser_click → ref: <resolution_button_ref>, element: "Resolution selector"
```

**Image Count** — Find the increment/decrement buttons (shows "1/4"):
- Click the "Increment" button to increase count (max 4)
- Click the "Decrement" button to decrease

**Character/Moodboard** — The right side panel shows "CHARACTER" with a "General" moodboard. Click "Change" to switch.

### Step 5: Generate

Take a screenshot to show the user the form state.

**ASK THE USER FOR CONFIRMATION** before clicking Generate — this uses credits/free generations.

Find and click the Generate button:
```
browser_click → ref: <generate_bu