← 返回 Skills 市场
Image Compressor
作者
ziji1224054593
· GitHub ↗
· v3.0.30
· MIT-0
150
总下载
0
收藏
1
当前安装
5
版本数
在 OpenClaw 中安装
/install image-compressor
功能描述
基于 `rv-image-optimize` 的图片压缩、上传与一体化压缩后上传 skill,支持 JPEG/PNG/WebP/AVIF、批量目录处理、FormData 接口配置与 JSON 结果汇总。Use when the user asks to compress images, upload optimi...
使用说明 (SKILL.md)
\r \r
Image Compressor\r
\r
基于 rv-image-optimize 的图片压缩、上传与格式转换 skill。 \r
An image compression, upload, and format-conversion skill powered by rv-image-optimize.\r
\r
核心功能 / Features\r
\r
- 压缩单张或批量图片 / Compress single images or whole folders\r
- 转换
jpeg、png、webp、avif/ Convert betweenjpeg,png,webp, andavif\r - 压缩时限制尺寸 / Resize while compressing with max width and height\r
- 通过
upload子命令按FormData配置上传文件 / Upload files through theuploadsubcommand withFormDataconfig\r - 通过
pipeline子命令一条命令完成“压缩 + 上传” / Complete "compress + upload" in one command with thepipelinesubcommand\r - 通过
chunkUpload配置执行分片上传 / 断点续传 / Run chunked and resumable uploads throughchunkUploadconfig\r - 输出结构化 JSON 结果 / Return structured JSON summaries for agents\r
- 默认安全输出到新目录 / Default to safe output in a new directory\r \r
何时使用 / When to use\r
\r 当用户提到以下需求时使用:\r \r
- “压缩图片”\r
- “图片太大”\r
- “减小文件大小”\r
- “转成 WebP / AVIF”\r
- “批量处理整个图片目录”\r
- “压缩时顺便缩小尺寸”\r
- “让 Agent / CLI 自动处理图片”\r
- “压缩后上传到接口”\r
- “按接口配置上传文件”\r
- “大文件分片上传”\r
- “断点续传”\r \r Use this skill when the user asks to:\r \r
- compress one or more images\r
- reduce image file size\r
- convert images to
webporavif\r - batch-process an image folder\r
- resize images while compressing\r
- run image optimization from an agent or CLI workflow\r
- upload files to an API through CLI\r
- compress files and immediately upload them\r
- perform chunked uploads for large files\r
- resume interrupted uploads\r \r
默认策略 / Default behavior\r
\r 除非用户明确要求修改原图,否则优先使用安全模式:\r \r
- 使用
--output-dir输出到新目录\r - 使用
--json方便 Agent 汇总结果\r - 默认不删除原图\r
- 默认不替换原图\r
- 大文件上传优先把
chunkUpload写进--config\r \r Preferred command:\r \r
rv-image-optimize "{input}" --output-dir "{outputDir}" --format webp --quality 82 --json\r
```\r
\r
如果全局命令不可用,可回退到:\r
\r
```bash\r
npx rv-image-optimize "{input}" --output-dir "{outputDir}" --format webp --quality 82 --json\r
```\r
\r
如果任务是“压缩后上传接口”,优先使用:\r
\r
```bash\r
rv-image-optimize pipeline "{input}" --format webp --quality 82 --config "{configPath}" --json\r
```\r
\r
如果任务是“已有文件直接上传接口”,优先使用:\r
\r
```bash\r
rv-image-optimize upload "{input}" --config "{configPath}" --json\r
```\r
\r
## 安全规则 / Safety rules\r
\r
以下选项只有在用户明确要求时才允许使用:\r
\r
- `--delete-original`\r
- `--replace-original`\r
\r
`--replace-original` 不能和以下参数一起使用:\r
\r
- `--output`\r
- `--output-dir`\r
- `--suffix`\r
\r
If the user wants to preserve source files, always choose `--output-dir`.\r
\r
上传相关安全约定:\r
\r
- 当前只保留 `FormData` 请求方式\r
- 复杂上传字段优先写进 `--config` JSON 文件\r
- 大文件 / 可恢复上传优先使用 `chunkUpload` 配置块\r
- `Authorization`、`Cookie`、`Content-Type` 优先用配置文件显式字段表达\r
- `Content-Type` 在 `FormData` 模式下通常建议留空,让运行时自动生成 boundary\r
\r
## 常用命令 / Common recipes\r
\r
### 单张图片压缩 / Compress a single image\r
\r
```bash\r
rv-image-optimize "./photo.jpg" --output-dir "./compressed" --format webp --quality 82 --json\r
```\r
\r
### 批量压缩目录 / Compress a folder\r
\r
```bash\r
rv-image-optimize "./images" --output-dir "./images-compressed" --format webp --quality 82 --json\r
```\r
\r
### 调整尺寸并压缩 / Resize and compress\r
\r
```bash\r
rv-image-optimize "./images" --output-dir "./images-compressed" --format webp --quality 82 --max-width 1920 --max-height 1080 --json\r
```\r
\r
### 压缩成功后删除原图 / Delete originals after success\r
\r
```bash\r
rv-image-optimize "./images" --output-dir "./images-compressed" --format webp --quality 82 --delete-original --json\r
```\r
\r
### 直接替换原图 / Replace originals in place\r
\r
```bash\r
rv-image-optimize "./images" --format webp --quality 82 --replace-original --json\r
```\r
\r
### 直接上传已有文件 / Upload existing files\r
\r
```bash\r
rv-image-optimize upload "./dist/demo.webp" --config "./upload.config.json" --json\r
```\r
\r
### 分片上传 / 断点续传 / Chunked resumable upload\r
\r
```bash\r
rv-image-optimize upload "./large-assets" --config "./upload.chunk.config.json" --timeout-ms 10000 --json\r
```\r
\r
### 压缩后直接上传 / Compress then upload\r
\r
```bash\r
rv-image-optimize pipeline "./images" --format webp --quality 82 --config "./upload.config.json" --json\r
```\r
\r
### 仅预览上传字段 / Preview upload fields only\r
\r
```bash\r
rv-image-optimize upload "./dist/demo.webp" --config "./upload.config.json" --preview-only --json\r
```\r
\r
## 汇报结果 / Reporting results\r
\r
使用 `--json` 时,建议汇总:\r
\r
- `total`\r
- `success`\r
- `failed`\r
- 输出目录或是否替换了原图\r
- 失败文件列表\r
- 上传模式下的 HTTP 状态码或预览字段\r
\r
## 推荐参数 / Recommended defaults\r
\r
- 默认格式:`webp`\r
- 默认质量:`82`\r
- 更小体积优先且接受更慢编码时:`avif`\r
- 上传任务默认优先使用 `--config`\r
- 大文件上传默认优先使用 `chunkUpload`\r
- 压缩后上传默认优先使用 `pipeline`\r
- 如需更多示例和参数说明,见 [reference.md](reference.md)\r
安全使用建议
This skill appears to do what it says: compress and optionally upload images using the rv-image-optimize CLI. Before installing or using it, verify you trust the rv-image-optimize npm package (inspect its repo and recent versions). Be careful when creating upload.config.json: do not put long-lived secrets (Authorization tokens, cookies) into configs you don't trust or that will be sent to unknown endpoints. Use the --preview-only option to inspect form fields before performing uploads, and prefer output-dir/default safe options rather than --replace-original or --delete-original unless the user explicitly consents. If you run the CLI via npx, remember npx will execute code from the registry — run in a controlled environment if you need stronger isolation.
功能分析
Type: OpenClaw Skill
Name: image-compressor
Version: 3.0.30
The skill provides image optimization and uploading capabilities via the `rv-image-optimize` package, but it includes high-risk features such as file deletion (`--delete-original`) and the ability to upload files to arbitrary remote URLs. Specifically, `SKILL.md` and `reference.md` instruct the agent on how to handle sensitive session data like `Authorization` tokens and `Cookie` headers during the upload process. While these capabilities are aligned with the stated purpose of the tool, they provide a direct mechanism for data exfiltration and unauthorized file modification if the agent is targeted by prompt injection.
能力评估
Purpose & Capability
Name/description match the declared requirements and install: the skill requires the rv-image-optimize CLI (or npx) and installs the rv-image-optimize npm package. Those are exactly what an image-compression + upload wrapper would need.
Instruction Scope
SKILL.md only instructs running the CLI (or via npx) and writing/using upload config files; it does not ask the agent to read arbitrary system files or hidden credentials. Note: upload instructions explicitly allow Authorization/Cookie fields in the config, meaning user-supplied secrets in those config files will be transmitted to the configured endpoints — this is expected for an upload feature but is a potential data-exfiltration vector if misused.
Install Mechanism
Install uses the npm package rv-image-optimize (standard node install). This is an expected, traceable registry install and not a direct download from an arbitrary URL or a code-extract step.
Credentials
The skill declares no required environment variables or credentials. While the docs show fields for Authorization/Cookie in upload config files, those are user-supplied for legitimate upload workflows and are not requested automatically by the skill.
Persistence & Privilege
always is false, the skill is user-invocable and does not request persistent system-wide privileges or modify other skills' configurations. Autonomous invocation is allowed by default but not escalated here.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install image-compressor - 安装完成后,直接呼叫该 Skill 的名称或使用
/image-compressor触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v3.0.30
- Added explicit support and instructions for chunked and resumable uploads using the chunkUpload config block.
- Updated recommended usage scenarios and default strategies for large file uploads and interrupted upload resumption.
- Expanded example commands to include chunked and resumable upload operations.
- Enhanced safety and config guidance for FormData and chunked upload configurations.
v3.0.22
Summary: Adds image upload and "compress then upload" capabilities.
- Added support for uploading files using the upload subcommand with FormData config.
- Introduced pipeline subcommand for streamlined "compress + upload" in one command.
- Updated skill description and usage guidance to include upload workflows.
- Expanded common command recipes and safety rules for upload scenarios.
v3.0.2
## 插件亮点 / Highlights
- 🚀 跨框架可用:同一个包同时覆盖 React 组件、Vue / Webpack / 原生 JS 工具函数,以及 Node / CLI 场景。 / Cross-framework support: one package covers React components, Vue / Webpack / vanilla JS utilities, and Node / CLI workflows.
- 🌐 跨语言调用友好:Java / Python / PHP 等后端可直接通过 CLI 复用压缩能力。 / Friendly for multi-language backends: Java / Python / PHP and other services can reuse the compression pipeline through the CLI.
- 🖼️ 图片优化能力完整:支持多 CDN 参数适配、自动格式选择、响应式图片、懒加载和渐进式加载。 / Complete image optimization features: multi-CDN URL adaptation, automatic format selection, responsive images, lazy loading, and progressive loading.
- ⚙️ 浏览器与服务端双压缩链路:既支持浏览器端压缩 / 无损压缩,也支持 `node-compress` 在 Node 环境原生处理图片。 / Dual compression paths for browser and server: supports browser-side compression / lossless compression and native Node processing through `node-compress`.
- 📦 支持静态图片打包压缩:可通过 `rv-image-optimize/vite-plugin` 在 `build` 后自动优化 `dist` 内静态图片。 / Static asset build optimization: use `rv-image-optimize/vite-plugin` to optimize static images in `dist` after build.
- 📤 上传链路可复用:提供 `upload-core` 和 `upload` 两层入口。 / Reusable upload pipeline: ships both `upload-core` and `upload` entry points.
- 💾 缓存体系完善:内置 IndexedDB + Worker 缓存能力,支持多库多表、自动过期和配额检测。 / Solid caching system: built-in IndexedDB + Worker caching with multi-database, multi-table, expiration, and quota checks.
- 🧩 `JS / TS` 都可直接接入:从 `3.x` 开始提供官方 `.d.ts`。 / Ready for both `JS / TS`: official `.d.ts` files are included starting from `3.x`.
- 🤖 Agent 集成友好:可直接接入 `Cursor`、`Claude Code`、skills 型 Agent,推荐通过 CLI + `--json` 调用。 / Agent-friendly integration: works well with `Cursor`, `Claude Code`, and skill-based agents, especially through the CLI with `--json`.
## Agent 集成亮点 / Agent Integration
- 支持 `Cursor`、`Claude Code`、skills 型 Agent 直接接入。 / Supports direct integration with `Cursor`, `Claude Code`, and skill-based agents.
- 推荐统一通过 `rv-image-optimize` CLI 调用,避免 Agent 临时拼接脚本。 / Using the `rv-image-optimize` CLI is recommended so agents do not have to assemble ad hoc scripts.
- 支持 `--json` 结构化输出,方便 Agent 稳定解析成功数、失败数、输出目录和压缩结果。 / Supports structured `--json` output so agents can reliably parse success counts, failures, output paths, and compression results.
- 既适合安全输出到新目录,也支持在用户明确授权时删除原图或替换原图。 / Works for safe output into a new directory and can also delete or replace originals when explicitly authorized.
- 详细接入方式、推荐提示词和 skill 规则见 [AGENT_INTEGRATION.md](./AGENT_INTEGRATION.md)。 / See [AGENT_INTEGRATION.md](./AGENT_INTEGRATION.md) for detailed integration guidance, prompt suggestions, and skill rules.
v1.0.1
No changes detected in this version.
- Version remains at 1.0.0 with no updates or modifications.
v1.0.0
- Initial release of the image-compressor skill.
- Supports compression and format conversion for JPEG, PNG, WebP, AVIF, and GIF images.
- Enables batch processing, adjustable compression quality, and dimension resizing.
- Provides CLI commands and Node.js API for integration.
- Includes React components for lazy and progressive image loading.
- Offers recommended quality settings for different use cases.
元数据
常见问题
Image Compressor 是什么?
基于 `rv-image-optimize` 的图片压缩、上传与一体化压缩后上传 skill,支持 JPEG/PNG/WebP/AVIF、批量目录处理、FormData 接口配置与 JSON 结果汇总。Use when the user asks to compress images, upload optimi... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 150 次。
如何安装 Image Compressor?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install image-compressor」即可一键安装,无需额外配置。
Image Compressor 是免费的吗?
是的,Image Compressor 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Image Compressor 支持哪些平台?
Image Compressor 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 Image Compressor?
由 ziji1224054593(@ziji1224054593)开发并维护,当前版本 v3.0.30。
推荐 Skills