Humanize Image
Security checks across malware telemetry and agentic risk
Overview
This skill is explicitly designed to hide AI-generated image provenance from detectors and can permanently strip metadata from original files.
Avoid installing this for normal use. Its central purpose is to remove AI provenance and bypass detectors, and its scripts can alter original files. If you must examine it for authorized research, use copies only, run it in an isolated environment, and preserve original images and metadata.
VirusTotal
66/66 vendors flagged this skill as clean.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The skill can help present AI-generated images as less detectable or more human-origin, which may mislead people or platforms and violate terms of service or laws around provenance/watermarks.
The workflow is measured by evading AI detectors and gives steps to improve evasion, showing a primary anti-provenance/anti-detection purpose.
**Success rate** = percentage bypassing Hive/Illuminarty/AI or Not detectors. ... If still detected, try: Increase strength ... Process twice
Do not use this for public posting, platform evasion, impersonation, or misleading content; only handle it in clearly authorized research or testing contexts.
Original image metadata such as camera data, timestamps, location, copyright, and provenance labels may be irreversibly removed, even when the user expects only a new output file.
The Python processor calls ExifTool with overwrite_original on the input path before saving the output, so original source metadata can be permanently erased.
subprocess.run(["exiftool", "-all=", "-overwrite_original", str(filepath)], ...) ... self.remove_metadata()
Run only on copies in a sandbox, and change the tool to copy inputs first and require explicit confirmation before modifying originals.
Installing the dependencies may require elevated privileges and trust in local package sources.
The skill relies on manual system and Python package installation; this is purpose-aligned for image processing but changes the local environment and is not captured by a registry install spec.
sudo apt install -y imagemagick libimage-exiftool-perl python3 python3-pip pip3 install Pillow numpy
Review dependency sources, avoid blindly running privileged commands, and prefer a contained environment such as a VM or container.