Image Quality Filter

Detect and filter out low-quality images by analyzing blur, brightness, and resolution to clean up image datasets efficiently.

Security Scan

The skill's code and instructions match its stated purpose: it analyzes local images for blur/brightness/resolution and can list, move, or delete them; it does not request credentials or perform network activity.

Purpose & Capability

Name/description, SKILL.md, and the included Python script are consistent: all required functionality (blur via Laplacian, brightness, resolution, batch actions) is implemented and no unrelated capabilities (cloud access, unrelated binaries) are requested.

Instruction Scope

Runtime instructions only invoke the included script on a user-specified directory. The script reads local image files and can list/move/delete them; it does not reference other system paths, environment variables, or external endpoints. Note: the delete action prompts for confirmation, while move does not.

Install Mechanism

No install spec is present (instruction-only), and the README's pip dependency list (Pillow, numpy, opencv-python) is proportional and expected for image analysis. Nothing is downloaded from arbitrary URLs.

Credentials

The skill declares no environment variables or credentials. The code does not access secrets or external service tokens.

Persistence & Privilege

The skill is not always-enabled and does not modify other skills or system-wide agent settings. It runs only when invoked and has no mechanisms for persistent background presence.

Assessment

This skill appears coherent and works locally on image files. Before using it: (1) run it in 'list' mode first to review results; (2) back up your dataset or test on a small subset before using 'move' or 'delete'; (3) note that blur detection requires OpenCV—if cv2 is not installed the script will skip blur checks; and (4) because move operations do not prompt, be cautious when specifying the output directory. There are no network calls or credential requests in the code.

Detect and filter out low-quality images including blurry, dark, too bright, or low resolution images. Use when user wants to clean up image datasets by removing poor quality images.

Features

• Blur Detection: Detect blurry images using Laplacian variance
• Brightness Analysis: Find too dark or too bright images
• Resolution Filter: Remove low-resolution images
• Quality Score: Compute overall quality score
• Batch Processing: Process large image folders
• Multiple Actions: List, delete, or move low-quality images

Usage

# Scan for low quality images
python scripts/quality_filter.py scan /path/to/images/

# Filter with custom thresholds
python scripts/quality_filter.py scan /path/to/images/ \
  --blur-threshold 100 \
  --min-resolution 640x480 \
  --min-brightness 30 \
  --max-brightness 220

# Delete low quality images
python scripts/quality_filter.py scan /path/to/images/ --action delete

Examples

$ python scripts/quality_filter.py scan ./images/

Scanning 150 images...
Analyzing quality...
Found 12 low-quality images:

[BLUR]   photo_blurry.jpg (score: 45)
[BLUR]   image_low.jpg (score: 62)
[DARK]   dark_photo.jpg (score: 38)
[BRIGHT] overexposed.jpg (score: 41)
[RES]    tiny_image.png (320x240)

Total: 12 low-quality images removed

Quality Criteria

Criterion	Threshold	Description
Blur	< 100	Laplacian variance (lower = blurrier)
Brightness	30-220	Out of range is poor
Resolution	> 640x480	Below minimum is low quality

Installation

pip install pillow numpy opencv-python

Options

• --blur-threshold: Blur threshold (default: 100)
• --min-resolution: Minimum resolution (default: 640x480)
• --min-brightness: Minimum brightness 0-255 (default: 30)
• --max-brightness: Maximum brightness 0-255 (default: 220)
• --action: What to do (list, delete, move)
• --output: Output folder for --action move