Professional Image Compression Tool
Paste the following prompt into your AI chat to install this skill:
Please follow https://skillhub.cn/install/skillhub.md to install @user_0bd8e437/zaq272.
About this skill
Problem
Image delivery often gets stuck between two constraints: visible quality and file size. Manually tuning quality, checking each output, and repeating the process across folders makes the workflow hard to reproduce. image-compression-zaq turns these trade-offs into a parameterized compression pipeline.
How it works
The skill is built on Python Pillow and focuses on a few practical operations:
- Automatic mode: tests q90/q80/q70/q60/q50, presents a comparison table, and recommends a more stable quality point.
- Target size: uses --target-size with binary search to keep output near a specified KB limit.
- Batch parallelism: handles directories with --batch, --recursive, and --jobs, which fits photo assets and page resources.
- Format and size control: converts among JPEG, PNG, WebP, BMP, and TIFF, and can resize by percentage or fixed dimensions with --resize.
- Structured output: --json emits machine-readable results for scripts, while --info shows dimensions, color mode, and EXIF.
Boundaries
This skill targets common bitmap compression and conversion, not vector graphics or complex image editing. It does not overwrite files by default; batch processing skips existing files by default; and the parallel job count should stay within the CPU core budget. For transparent assets, WebP is usually a better fit than JPEG, while lossless PNG compression can be enabled with --lossless.
Use Cases
- Before launch, compress a page-image folder to under 200 KB and get a recommended quality.
- When organizing event assets, convert transparent PNG files to WebP and inspect image info.
- When scripting screenshot processing, read `--json` compression results and record file sizes.
- When exporting report images, resize to fixed dimensions and preserve EXIF metadata.
Best For
- Front-end engineers who need web assets compressed to a target KB size
- Operations staff who organize event images and keep formats consistent
- Automation engineers who batch-process screenshots and emit JSON results
- Design assistants who maintain product galleries and convert to WebP
Related Skills
Uses step-by-step choices to confirm business, palette, and layout, then exports an editable .drawio architecture diagram.
An agent skill for the Miaoyin AI music REST API that supports SUNO/Mureka-based song generation, continuation, cover, video, WAV, and stem splitting.
Distills text, files, or images into structured knowledge and generates knowledge-card prompts for drawing tools, supporting 3:4 portrait or 16:9 landscape outputs.
Capture full-page, viewport, or element screenshots with Playwright while handling lazy loading, internal scroll, mobile layouts, and Linux CJK fonts.