A structural diagnostic tool that measures compositional coordinates in AI-generated images to identify and correct model-specific biases.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install vtl-image-analysis
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 vtl-image-analysis 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 VTL Image Analysis?
VTL Image Analysis is a sophisticated diagnostic skill designed for developers and artists who need to quantify the structural quality of AI-generated imagery. By utilizing the Visual Thinking Lens (VTL) framework, this skill identifies "default-mode" behaviors—such as the tendency for models to center-lock subjects or collapse compositions radially—which often result in generic or flat visuals. This skill provides a deterministic way to move beyond subjective critiques and into data-driven image optimization using Openclaw Skills.
The tool functions as a bridge between raw image generation and professional-grade art direction. It calculates five specific structural coordinates and, when necessary, utilizes a rule-based engine to generate re-prompts that push the AI model out of its comfort zone. This ensures that the generated content reflects true authorship rather than model prior behavior.
VTL Image Analysis Use Cases
- Diagnosing why AI-generated images feel generic, flat, or boring.
- Automating quality control for high-volume image generation workflows.
- Generating targeted re-prompts to break a model's reliance on center-weighted compositions.
- Benchmarking the structural variety and compositional health of different LLM and diffusion model outputs.
How VTL Image Analysis Works
- The skill executes a Python-based probe script to analyze the structural signal and edges within a target image.
- A strict refusal gate evaluates the measurement's validity; if the signal is too sparse, the analysis stops to prevent false diagnostics.
- The system extracts five core metrics: Placement (delta), Void (r_v), Packing (rho_r), Radial (dRC), and Tension (k_var).
- If specific bias flags like CENTER_LOCK are triggered, the skill references a local configuration file to generate up to three optimized re-prompt variants.
VTL Image Analysis Setup
To utilize this skill within the Openclaw Skills ecosystem, ensure you have the required Python dependencies installed:
pip install numpy opencv-python-headless scikit-image scipy pyyaml
The tool requires python3 to be available in your system path to execute the diagnostic scripts.
VTL Image Analysis Data Schema & Taxonomy
The skill produces a structured JSON output representing the image's compositional coordinates:
| Coordinate | Key | Purpose |
|---|---|---|
| Placement | delta_x, delta_y |
Measures the mass centroid offset from the frame center. |
| Void | r_v |
Quantifies the amount of empty or negative space. |
| Packing | rho_r |
Determines the density and distribution of visual elements. |
| Radial | dRC |
Identifies if the composition is collapsing toward or expanding from the center. |
| Tension | k_var |
Measures the variation in structural complexity across the frame. |
VTL Image Analysis Advanced Features
- Deterministic Operator Logic: Uses
operators.yamlto ensure re-prompts are rule-based rather than hallucinated by the AI. - Non-negotiable Refusal Gate: Prevents analysis of low-contrast or low-signal images where measurements would be unreliable.
- Radial Eligibility Detection: Automatically labels "dual-center" compositions where mass is too close to the center for standard radial measurement.
- Multi-Variant Prompting: Provides a primary recommended prompt and several alternatives based on specific compositional flags.
SKILL.md
Loading
Related Openclaw Skills
Threads API Integration
A professional CLI tool for managing Threads profiles and automating content publication via the Graph API.
Pinterest API v5 Integration
A comprehensive Pinterest API v5 integration for automating pin creation, board management, and performance tracking.
Obsidian Knowledge Base
Connect your AI agent to your Obsidian vault for seamless note creation, fuzzy searching, and automated knowledge management.
Obsidian Knowledge Base
A comprehensive AI skill for managing Obsidian vaults through intelligent search, automated organization, and granular content editing.
Catch My Skill
An automated utility to synchronize and track version differences for your local Openclaw Skills against GitHub and ClawHub repositories.
Class Seven Multi-Agent Development Team
Class Seven orchestrates a specialized multi-agent development team to handle complex software tasks from planning to deployment.
Featured*
