An AI-powered image processing skill that resizes and extends visual assets into multiple platform-specific dimensions while ensuring element integrity.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install image-assets-resize
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install image-assets-resize using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Image Assets Resize?
Image Assets Resize is a professional-grade tool designed for marketing and design workflows within the Openclaw Skills ecosystem. It leverages advanced vision models like GPT Image 2 and Nano Banana 2 to take a single Key Visual (KV) or poster and intelligently adapt it into various sizes for social media, e-commerce, and web platforms.
Unlike traditional cropping tools, this skill uses AI edit modes to natively understand image content, allowing it to extend backgrounds and reposition elements without manual intervention. It eliminates the need for tedious manual resizing, ensuring that logos, characters, and text remain visually balanced and intact across every output format.
Image Assets Resize Use Cases
- Automatically generating cover images for WeChat, Douyin, Xiaohongshu, and Bilibili from a single reference poster.
- Extending the background of a horizontal Key Visual to fit vertical 9:16 story formats for Instagram or Facebook.
- Batch producing clean versions of marketing materials by removing logos or specific text elements using AI instructions.
- Creating custom-dimension assets for web banners and e-commerce product pages with precise pixel requirements.
- High-volume creative production where AI self-verification is needed to check for text clipping or element distortion.
How Image Assets Resize Works
- The user provides a reference image path and specifies target platforms or custom dimensions.
- The system queries its platform specification database to determine the exact required pixel dimensions and aspect ratios.
- It calculates a valid generation size (multiples of 16 and at least 1MP) to satisfy the requirements of the underlying AI vision models.
- A specialized prompt is sent to the backend (gpt-image-2 or nano-banana) to generate or edit the image based on the reference.
- The tool performs a high-precision downscale (usually <1% change) to match the exact target dimensions without visual loss.
- If the verify flag is enabled, an AI agent inspects the final result for visual quality and element integrity.
- Files are saved using a standardized naming convention and, if in cloud mode, returned as base64 strings for immediate display.
Image Assets Resize Setup
To get started with this skill in your Openclaw Skills environment, ensure you have python3 and the necessary API keys.
# Initialize configuration and select your preferred AI model
python scripts/generate_image.py --setup
# Set your environment variables
export OPENAI_API_KEY='your_key_here'
export GPT_IMAGE2_API_KEY='your_key_here'
Image Assets Resize Data Schema & Taxonomy
The skill organizes data within a dedicated configuration directory and standardized output folders:
| Type | Path / Format | Description |
|---|---|---|
| Config | ~/.content-marketing/config.json |
Stores API endpoints and active backend model. |
| Specs | ~/.content-marketing/custom_specs.json |
Stores user-defined platform dimensions. |
| Output | output/{scene}_{platform}_{WxH}_v{version}.png |
Final generated image assets. |
| Results | JSON Object | Returns success status, local paths, and base64 strings for cloud integration. |
Image Assets Resize Advanced Features
- Multi-platform Batch Processing: Generate assets for dozens of social networks with a single command.
- AI Self-Verification: Use GPT-4o to automatically detect and flag layout errors or cut-off text.
- Custom Dimension Learning: Define new platform specs on the fly which the system remembers for future use.
- Extra Prompt Injection: Add or remove specific elements (like 'Add a Red Button' or 'No Logo') during the resize process.
- Brand Language Feedback: Record ratings and feedback on generations to help the system adapt to specific brand design languages.
SKILL.md
Loading
Related Openclaw Skills
Notion Integration
A comprehensive toolset for managing Notion workspaces via API, enabling automated database operations and page content management.
Automation Workflow
A robust framework for creating and managing automated multi-step business workflows and periodic data synchronization tasks.
Auto-Spec
A spec-driven development assistant that helps developers align intent through behavioral contracts before coding or by reverse-engineering existing logic.
Data Scraper
A powerful data extraction tool that turns unstructured website content into clean, structured datasets for analysis.
RSS Aggregator
A powerful utility to parse, filter, and aggregate multiple RSS/Atom feeds into structured JSON, HTML, or Markdown formats.
Telegram Bot
Create and manage Telegram bots for automated notifications, customer support, and interactive group management.
Featured*
