A sophisticated visual cortex for text-based agents that enables image analysis and description via the MoltShell M2M Marketplace.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install moltshell-vision
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 moltshell-vision 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 MoltShell Vision Engine?
The MoltShell Vision Engine is a critical upgrade for developers using Openclaw Skills who need to overcome the visual limitations of text-only AI models. While standard agents often fail when encountering non-textual data like infographics or UI screenshots, this skill provides a robust visual-to-text bridge. It securely routes image data to advanced Vision-Language Models, allowing your agent to understand and act upon visual information as if it were native text.
By integrating this skill, your bots gain the ability to navigate image-heavy websites, interpret complex dashboards, and process user-uploaded media. This adds a layer of multi-modal capability to your existing Openclaw Skills ecosystem without requiring complex local model hosting or heavy infrastructure management.
MoltShell Vision Engine Use Cases
- Processing and describing web-scraped infographics for research agents.
- Analyzing dashboard screenshots to verify UI states in automated testing workflows.
- Interpreting user-uploaded images in customer support or chatbot scenarios.
- Extracting text and context from visual media within data pipelines.
How MoltShell Vision Engine Works
- The agent identifies an image URL and a specific visual query (prompt) that needs to be resolved.
- The
moltshell_visiontool is invoked, sending the image metadata to the MoltShell M2M Marketplace. - A high-performance Vision-Language Model (VLM) processes the image based on the provided instructions.
- The model generates a descriptive text response representing the visual content.
- The text description is returned to the Openclaw Skills agent, which then incorporates this data into its reasoning loop.
MoltShell Vision Engine Setup
This skill includes a built-in sandbox for immediate testing. To move to production within your Openclaw Skills setup, follow these steps:
- Sign up and generate an API key at https://moltshell.xyz.
- Configure your environment variable to enable production access:
export MOLTSHELL_API_KEY=sk_molt_your_key_here
- The skill will automatically switch from the sandbox to your dedicated account once the environment variable is detected.
MoltShell Vision Engine Data Schema & Taxonomy
The skill facilitates structured visual reasoning using the following schema:
| Parameter | Type | Required | Description |
|---|---|---|---|
image_url |
string | Yes | The publicly accessible link to the image file. |
prompt |
string | Yes | The natural language instruction for the vision model. |
output |
string | N/A | A detailed text description of the image content. |
MoltShell Vision Engine Advanced Features
- Zero-config public sandbox providing $0.25 of free shadow-compute for instant developer testing.
- Seamless environment variable integration for scaling from prototype to production.
- Access to the MoltShell M2M Marketplace ecosystem for diverse model selection.
- Low-latency image processing designed specifically for Openclaw Skills agent loops.
SKILL.md
Loading
Related Openclaw Skills
PurposeBot
PurposeBot is a comprehensive agentic commerce layer that enables AI agents to discover tools, execute payments via Stripe or USDC, and build verifiable reputation.
Radarr and Sonarr Integration
A powerful integration for OpenClaw that automates movie and TV series management via Radarr and Sonarr servers.
Wolt CLI
A comprehensive command-line tool for browsing Wolt venues, managing carts, and automating delivery profile tasks through structured data.
Wolt CLI
A robust command-line interface for interacting with the Wolt marketplace to browse venues, manage carts, and handle profile tasks.
Forge
Forge is a repair-inspect loop orchestrator that automates code fixes with independent verification and dependency-aware parallel execution.
PulpMiner
PulpMiner is an AI-powered tool that converts any webpage into clean, structured JSON data using custom schemas and LLM-driven extraction.
Featured*
