An advanced image manipulation skill for AI agents to edit, modify, and replace elements in images while preserving pixel-perfect original resolution.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install image-editor
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-editor 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 Editor?
The Image Editor skill empowers AI agents to perform complex pixel-level image manipulations directly using Python's Pillow (PIL) library. Instead of relying on lossy external APIs or resizing utilities, this capability allows agents to handle advanced modifications, such as text replacement and element extraction, while maintaining the file's original resolution. As a vital addition to your collection of Openclaw Skills, this tool guarantees precise, high-fidelity operations on standard image formats.
By integrating this skill, agents gain the intelligence to locate coordinates, draw or fill shapes, overlay typography, and dynamically modify graphics. Crucially, the workflow embeds a mandatory feedback loop using vision models to ensure accuracy before final output delivery.
Image Editor Use Cases
- Replacing text or labels inside structural schemas, screenshots, or localized banner graphics.
- Removing unwanted visual elements or artifacts by dynamically masking and filling background textures.
- Adding text annotations, labels, or watermarks to high-resolution diagrams without altering their scale or quality.
- Performing precise bounding box calculations and visual QA comparisons using vision language models.
How Image Editor Works
- Analyze Original Image: Read the source file's metadata (dimensions, color mode) using Python script-based PIL diagnostics.
- Identify Target Coordinates: Use the agent's vision tool to locate exact pixel boundaries (x_start, y_start, x_end, y_end) for the edits.
- Edit via PIL Direct Manipulation: Execute precise drawing operations, such as laying down background fill and writing crisp text layers using native system fonts.
- Verify Output via Vision Model: Run an automated quality check by requesting a comparison of the modified image against the original to confirm resolution fidelity and layout accuracy.
- Deliver Output: Safely deliver the processed file only after the visual verification pass succeeds.
Image Editor Setup
To run this skill, ensure that python3 and Pillow (PIL) are installed in your environment.
# Ensure Python 3 is installed
python3 --version
# Install the required PIL library (Pillow)
pip install Pillow
This setup works seamlessly out of the box with Openclaw Skills configurations requiring Python binary execution.
Image Editor Data Schema & Taxonomy
The Image Editor skill relies on structural parameters and system font paths to function correctly.
| Parameter / Resource | Description | Example Path / Value |
|---|---|---|
| Input Formats | Standard visual file formats | JPEG, PNG |
| macOS Native Font | Default system font library path | /System/Library/Fonts/STHeiti Light.ttc |
| Quality Control | Quality compression ratio for output JPEG | 95 |
| Coordinates | Absolute bounds for rendering canvas | (x_start, y_start, x_end, y_end) |
Image Editor Advanced Features
- Multi-Agent Quality Check: Employs an integrated vision feedback loop that tasks a separate vision-language model to review and confirm visual correctness.
- Dynamic Font Scaling: Programmatic coordinate mapping ensures text replacement layers match the exact canvas sizes and original styles perfectly.
- Pixel-Level Integrity Protection: Enforces a strict non-destructive rule preventing any automatic downscaling, resizing, or lossy web-based fetching routines.
SKILL.md
Loading
Related Openclaw Skills
AgentPMT No-Account AgentAddress & x402
An integration profile under the Openclaw Skills ecosystem that lets autonomous agents utilize AgentPMT tools and premium workflows without an account, using cryptographic wallets, scoped credits, or direct x402 micropayments.
The Storm of Steel: Ernst Jünger WWI Memoir Assistant
An interactive AI agent skill that lets you explore and analyze Ernst Jünger's classic WWI memoir, The Storm of Steel, for deep historical insights and leadership lessons.
Install AgentPMT MCP Server
A specialized skill to connect MCP-compatible AI agents, IDEs, and programmatic clients to the AgentPMT tool marketplace with budget enforcement.
Waking the Tiger: Somatic Experiencing AI Agent Skill
An AI-driven guide to Peter Levine's Somatic Experiencing, enabling users to explore body-based trauma healing, somatic techniques, and nervous system regulation.
The Hero with a Thousand Faces
An AI agent skill based on Joseph Campbell's masterwork that maps universal mythological structures, the monomyth lifecycle, and psychological archetypes into actionable intent routing.
The Education of an Idealist Skill
An AI agent skill translating Samantha Power's memoir into executable mental models for risk management, strategic empathy, and ethical leadership.
Featured*
