yolo-vision-tools
PassAudited by ClawScan on May 1, 2026.
Overview
The provided artifacts look like a legitimate YOLO computer-vision helper, with expected cautions around installing Ultralytics, running local diagnostics, and using selected media or webcam input.
This skill appears safe for its stated purpose. Before installing, use a virtual environment, verify the Ultralytics package/version you want, run local diagnostic scripts only when needed, and be deliberate about webcam or private image/video inputs and saved outputs.
Findings (3)
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.
Installing or updating the dependency changes the local Python environment and executes package-provided code during normal use.
The setup guidance installs or updates an external Python package without a version pin. This is expected for a YOLO skill, but it means the user trusts the package source and current package version.
pip install -U ultralytics
Install in a virtual environment and pin or verify the Ultralytics version if reproducibility or supply-chain control matters.
Running the helper may reveal local environment details such as Python paths, platform information, GPU details, and installed package versions.
The environment-check helper runs local shell commands for diagnostics. This is consistent with checking Python/CUDA/YOLO setup, but users should understand that running the helper executes commands on their machine.
result = subprocess.run(cmd, shell=True, capture_output=True, text=True, check=True)
Run diagnostic scripts only when needed, preferably in the intended project environment, and avoid sharing the output publicly if it contains local paths or system details.
If invoked with webcam or private media, the agent may process sensitive visual content and save annotated results locally.
The skill documents optional webcam processing and saving analysis outputs. This is normal for computer-vision work, but it touches local media/device input.
results = model(0, show=True) # webcam with display → saved to yolo-vision/outputs/videos/
Use the skill only with media sources you intend to analyze, and review where outputs are saved before sharing them.