image-nuke

What This Skill Does

Image Nuke is an industrial-grade privacy tool designed to sanitize image files for total anonymity. Unlike standard tools that simply delete metadata, Image Nuke employs a multi-layered approach to ensure that your files are forensically untraceable and resistant to reverse image search engines. The tool strips all EXIF, GPS, XMP, IPTC, and ICC data while performing destructive re-encoding. To defeat fingerprinting, it injects sub-pixel Gaussian noise, rotates color hues, and performs random micro-crops that alter the image dimensions by 1-3 pixels. This renders the file visually identical to the human eye but mathematically distinct from the original, effectively destroying the perceptual hash used by search engines and social media platforms to track or identify images across the internet.

Installation

To integrate this utility into your OpenClaw agent, use the package manager included in the framework. Run the following command in your terminal:

clawhub install openclaw/skills/skills/cassh100k/image-nuke

Ensure you have the necessary system dependencies installed: Python 3, Pillow for image processing, and NumPy for the high-precision noise injection calculations. You can install the required Python libraries using pip install Pillow numpy.

Use Cases

Image Nuke is ideal for journalists, whistleblowers, and privacy-conscious users who need to share media without risking location disclosure via GPS data or facial recognition tracking. It is also highly effective for social media users who wish to prevent their personal images from being scraped for training datasets or cross-platform identity matching. By normalizing files into a clean JPEG format with randomized noise, users can ensure their visual content carries no digital signature.

Example Prompts

1. "Image Nuke, process my vacation photos in ./uploads/ and save the cleaned versions to ./safe_output/ using noise level 4 to ensure they are fully scrubbed."
2. "I need to upload this image to a public forum. Please run image-nuke on portrait.jpg with a quality setting of 85 and maximum noise to prevent reverse searching."
3. "Sanitize the image located at ./work/doc_scan.png using the standard noise level 3 and output the result as a clean jpeg file."

Tips & Limitations

• Output Format: Please note that Image Nuke strictly converts all input files to JPEG. If you require PNG transparency, this tool is not suitable, as the re-encoding process drops alpha channels.
• Integrity: The original file is never overwritten, making this safe for bulk directory operations. However, ensure you have sufficient disk space when batch processing high-resolution images, as the process creates entirely new copies.
• Quality: Higher noise levels (Level 5) offer the best protection but may slightly degrade image clarity. Use Level 2 or 3 for standard web sharing where balance between anonymity and visual quality is required.