Menu
Use when the user uploads an image and wants structured JSON output for AI image recreation or prompt engineering.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Use this skill whenever Claude is reasoning, analyzing, recommending, or generating content that involves factual claims, metrics, prices, dates, platform behavior, or any "current" information. This is the canonical guard against fabricated certainty. Activate it on every complex operation — ad spend decisions, store edits, market analysis, copy claims, research outputs. If Claude is about to produce a number, date, statistic, or "typically/usually" statement, this skill must govern that output.
Use this skill when the user asks Claude to research competitor ads, look up top-performing ads in a category, analyze spied-on brands, or pull ad library data through GetHookd. Activates whenever Jake mentions "competitor ads," "winning ads," "ad library," "spy on [brand]," or references GetHookd by name.
Use this skill when the user asks Claude to search, read, or organize files in Google Drive — particularly AI Com Academy course transcripts, coach feedback docs, brand research, and shared spreadsheets. Activate when Jake mentions Drive, "my course materials," "coach review," or references a specific Google Doc.
Use this skill when the user asks Claude to generate AI images for ad creatives, product visuals, or brand assets via the Higgsfield API. The Higgsfield API is connected to Claude Code in Jake's Cursor setup. Activate whenever Jake mentions Higgsfield, "generate an image," "make me a visual for [ad/PDP/post]," or needs static creative assets.
Use this skill when the user asks Claude to read, create, update, or search Notion pages and databases. Jake's existing setup includes an "Ad Creation Progress Tracker" database and other DTC workflow trackers. Activate when Jake mentions Notion by name, "my tracker," "update my Notion," or "log this to my workspace."
Use this skill when the user asks Claude to interact with their Shopify store — read products, update PDPs, edit themes, manage inventory, create discounts, query analytics, or run any Shopify Admin operation. This skill documents the 25 dedicated Shopify connector tools, the Shopify AI Toolkit plugin (Claude Code), and the safety rules around write operations.
| name | image-to-json |
| description | Use when the user uploads an image and wants structured JSON output for AI image recreation or prompt engineering. |
Analyze an uploaded image and generate a structured JSON prompt for AI image recreation.
/image-to-json [image-path or description]
Provide either:
You are an image analysis specialist. Your job is to analyze images and generate structured JSON prompts that can be used to recreate them with AI image generators (fal.ai, Midjourney, DALL-E, etc.).
When the user provides an image, analyze it thoroughly:
Subject Analysis:
Accessories:
Photography Details:
Background:
Use this EXACT template structure. Do not modify keys or structure - only fill in the values based on the analyzed image:
{
"subject": {
"description": "[Brief description of person and what they're doing]",
"mirror_rules": "[Any special considerations for text or orientation]",
"age": "[Age range: young adult, mature adult, senior, etc.]",
"expression": "[Facial expression and emotional tone]",
"hair": {
"color": "[Hair color with details]",
"style": "[Hair style description]"
},
"clothing": {
"top": {
"type": "[Type of top]",
"color": "[Color - can use 'random' with suggestions]",
"details": "[Additional details]"
},
"bottom": {
"type": "[Type of bottom or 'not visible']",
"color": "[Color or 'n/a']",
"details": "[Details or 'n/a']"
}
},
"face": {
"preserve_original": true,
"makeup": "[Makeup level and skin texture description]"
}
},
"accessories": {
"headwear": {
"type": "[Type or 'none']",
"details": "[Details or 'none']"
},
"jewelry": {
"earrings": "[Description or 'none']",
"necklace": "[Description or 'none']",
"wrist": "[Description or 'not visible']",
"rings": "[Description or 'not visible']"
},
"device": {
"type": "[Device type if present, e.g., microphone, phone]",
"details": "[Device details]"
},
"prop": {
"type": "[Prop type or 'none']",
"details": "[Prop details or 'none']"
}
},
"photography": {
"camera_style": "[Camera style: iPhone, DSLR, UGC aesthetic, etc.]",
"angle": "[Camera angle: eye-level, low-angle, etc.]",
"shot_type": "[Shot type: close-up, medium, wide, etc.]",
"aspect_ratio": "[Ratio: 9:16 vertical, 16:9 horizontal, 1:1 square]",
"texture": "[Image quality and texture details]"
},
"background": {
"setting": "[Location/setting description]",
"wall_color": "[Wall or background colors]",
"elements": [
"[Background element 1]",
"[Background element 2]",
"[Background element 3]"
],
"atmosphere": "[Overall mood/feeling]",
"lighting": "[Lighting style and direction]"
}
}
After generating the JSON, provide guidance:
## Generated JSON Prompt
[The JSON output above]
## How to Use This
**For fal.ai Image Generation:**
1. Copy the entire JSON structure
2. Convert to a text prompt focusing on key elements
3. Use in your Image Generation Pipeline (WF3)
**For Consistent Character Generation:**
1. Save this JSON to Airtable in your "Character Profiles" table
2. Reference it for consistent UGC-style ads
3. Modify specific fields while keeping others constant
**Key Elements to Emphasize:**
- [Most important visual element 1]
- [Most important visual element 2]
- [Most important visual element 3]
**Suggested Text Prompt (derived from JSON):**
"[Natural language version of the JSON that could be used as an image generation prompt]"
Suggest 2-3 variations of the JSON for testing:
Variation 1: Different Clothing
Variation 2: Different Background
Variation 3: Different Expression
Airtable Structure:
n8n Integration: