Token导航 LogoToken导航TokenDH.com
图像处理权限需确认github未标认证来源可访问许可证需确认审计通过

image-optimizer图像优化器

Agent Skill

用于辅助图像生成、图片编辑、视觉素材处理或图像模型工作流。它适合让 Agent 根据文本生成图片、处理背景、整理视觉提示词或调用相关图像工具。使用时需要确认输入图片、版权来源、输出格式和模型限制;涉及人物、品牌、商品或公开展示素材时,应额外核对授权、真实性和内容合规边界。

总安装

461

周安装

19

GitHub Stars

4

下载量

150
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:image-optimizer(图像优化器)
来源仓库:https://github.com/insight68/skills
仓库路径:skills/image-optimizer
安装命令:
npx skills add https://github.com/insight68/skills --skill image-optimizer
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/insight68/skills --skill image-optimizer

简介

自动化图像优化流水线,支持预设规则执行。

  • 适用于CMS后台、上传审核与定期维护任务。
  • 可集成到CI/CD流程中实现持续优化。
  • 注意处理耗时与服务器负载平衡。image-optimizer 属于图像处理类 Skill,可作为该场景下的辅助能力补充。
  • 建议保留原始文件用于追溯与重新处理。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Image Optimizer

Image optimization toolkit using Sharp for high-performance WebP conversion, thumbnail generation, and intelligent background removal.

Quick Start

All scripts are in scripts/. Use pnpm to run them:

# Basic WebP conversion (large images)
node scripts/optimize-images.js

# Small images (higher quality)
node scripts/optimize-small-images.js

# Advanced with custom config
node scripts/optimize-images-advanced.js

Thumbnail Generation

Generate thumbnails with specific dimensions or aspect ratios:

# Square thumbnails (128x128)
pnpm optimize:128

# Square thumbnails (256x256)
pnpm optimize:256

# 16:9 aspect ratio (640x360 default)
pnpm optimize:16-9

# Process specific files
pnpm optimize:128 public/image.png public/logo.png

# Custom options
node scripts/optimize-thumbnails.js --size=128 --dir=public/icons

Aspect Ratio Thumbnails

Generate thumbnails with preset aspect ratios:

# 16:9 (widescreen, YouTube)
pnpm optimize:16-9 --size=small    # 640x360
pnpm optimize:16-9 --size=medium   # 800x450
pnpm optimize:16-9 --size=large    # 1920x1080

# 4:3 (traditional, iPad)
pnpm optimize:4-3 --size=medium    # 800x600

# 1:1 (square, Instagram)
pnpm optimize:1-1 --size=large     # 1080x1080

# Custom dimensions
node scripts/optimize-aspect-ratio.js --width=1280 --height=720

Background Removal

Remove white backgrounds with smart edge smoothing:

# Basic white to transparent
pnpm optimize:128 public/product.png --transparency=white-to-transparent

# High quality for product photos
pnpm optimize:256 public/product.jpg \
  --transparency=white-to-transparent \
  --threshold=250 \
  --quality=90

# Custom threshold for off-white backgrounds
pnpm optimize:128 public/icon.png --transparency=white-to-transparent --threshold=230

Configuration

WebP Quality

  • 85-90: Product photos, high-quality images
  • 75-85: Standard web images (default)
  • 60-75: Thumbnails, non-critical images

Fit Strategies

  • cover: Fill entire area (recommended, exact dimensions)
  • inside: Maintain aspect ratio, may not fill completely
  • fill: Force stretch (distorts, not recommended)

Crop Position

  • entropy: Smart crop, keeps subject (recommended)
  • center: Center crop
  • attention: AI attention-based (requires Sharp 0.29+)
  • top, bottom, left, right: Edge-aligned

Transparency Modes

  • auto: Auto-detect based on input (PNG/WebP preserve, others remove)
  • preserve: Keep existing transparency
  • remove: Remove transparency, fill with background color
  • white-to-transparent: Smart white background removal
  • add: Add transparency with alpha level

Background Removal Options

  • --threshold: White threshold (0-255, default 240). Use 230-245 for off-white, 250-255 for pure white
  • --smooth-edges: Enable edge smoothing (default true)
  • --supersample: 2x supersampling for smooth edges (default true)
  • --gaussian-smoothing: Use 5x5 Gaussian kernel (default true, false=3x3 simple)
  • --smoothing-iterations: Edge smoothing passes (default 1, max 2)

Common Use Cases

E-commerce Product Photos

# Remove white background, high quality
pnpm optimize:256 public/products/*.jpg \
  --transparency=white-to-transparent \
  --threshold=250 \
  --quality=90 \
  --supersample=true \
  --gaussian-smoothing=true

Blog Thumbnails (16:9)

# Generate 16:9 thumbnails for blog posts
pnpm optimize:16-9 --size=medium --dir=public/blog-images

Social Media Avatars (1:1)

# Square thumbnails for profiles
pnpm optimize:1-1 --size=large --dir=public/avatars

Icon Optimization (128x128)

# Small square icons
pnpm optimize:128 public/icons/logo.png --transparency=preserve

Processing Directories

Batch process entire directories:

# Process all images in a directory
node scripts/optimize-thumbnails.js \
  --dir=public/uploads \
  --size=256 \
  --output=public/thumbnails

# With background removal
node scripts/optimize-thumbnails.js \
  --dir=public/products \
  --size=512 \
  --transparency=white-to-transparent \
  --threshold=245

Tips

  1. Use supersampling for product photos to eliminate jagged edges
  2. Higher quality (90+) for images that will be further edited
  3. Standard quality (75) for final web images
  4. Entropy position preserves important image content
  5. Test thresholds on sample images before batch processing
  6. Lower threshold (230-240) for off-white backgrounds
  7. Higher threshold (250-255) for pure white backgrounds

Dependencies

All scripts require:

pnpm add sharp

Sharp is a high-performance Node.js image processing library.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

32.59%
按下载量换算49

Claude

29.75%
按下载量换算45

Cursor

19.44%
按下载量换算29

Gemini CLI

10.06%
按下载量换算15

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

权限需确认

当前来源未能明确判断权限范围,默认进入异常复核队列。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills