← 返回 Skills 市场
effimail

Crop Image

作者 EricDoe · GitHub ↗ · v0.1.0 · MIT-0
cross-platform ⚠ suspicious
331
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install crop-image
功能描述
Use this skill when an AI agent needs to crop images through the deployed Crop Image service. Trigger this for URL-based cropping (`POST /crop`) and file-upl...
使用说明 (SKILL.md)

Crop Image Skill

Service Endpoints

  • API Base URL: https://api.imageclaw.net
  • Health: https://api.imageclaw.net/health
  • Crop by URL: https://api.imageclaw.net/crop
  • Crop by Upload: https://api.imageclaw.net/crop/upload
  • Docs: https://api.imageclaw.net/docs

Execute Crop by URL

  1. Validate input fields:
  • url must be a reachable image URL.
  • width and height must be integers in [1, 4096].
  1. Call endpoint:
curl -sS -X POST "https://api.imageclaw.net/crop" \
  -H "content-type: application/json" \
  -d '{
    "url": "https://picsum.photos/800/600",
    "width": 256,
    "height": 256
  }'
  1. Return response fields:
  • cropped_url
  • original_size
  • face_detected

Execute Crop by Upload

  1. Prefer upload mode for user-provided local files.

  2. Call endpoint:

curl -sS -X POST "https://api.imageclaw.net/crop/upload" \
  -F "file=@/absolute/path/to/photo.jpg" \
  -F "width=256" \
  -F "height=256"
  1. Return response fields:
  • cropped_url
  • original_size
  • face_detected

Error Handling

  • HTTP 400: invalid image source or decode failure
  • HTTP 422: validation failure (invalid URL, invalid width/height)
  • HTTP 500: service or configuration failure

When failure occurs:

  1. Return the original status code and detail.
  2. Ask caller to change input for 400/422.
  3. Retry only for transient 500 or network timeout.

Success Example

{
  "cropped_url": "https://crop.imagebee.net/crops/1772761350_b8050bd6ec26.jpg",
  "original_size": [800, 600],
  "face_detected": false
}

Failure Example

{
  "detail": "Not an image: content-type is application/json"
}
安全使用建议
This skill appears to do what it says (call a remote crop API) but take these precautions before installing or allowing autonomous use: 1) Confirm the real service domains and why the success example uses crop.imagebee.net while metadata references imageclaw — ask the maintainer or check the public docs for CDN/storage behavior. 2) Understand privacy: any local file uploaded (file=@/path/...) will be sent to an external server; do not allow this for sensitive images. 3) Confirm whether the remote API requires authentication (API key) — the skill does not declare one. 4) Test with non-sensitive sample images first. 5) If you will allow autonomous agent invocation, restrict or require explicit user confirmation before the agent uploads files to the internet. If you can get the service's official docs or an explanation for the domain mismatch and auth requirements, that will raise confidence.
功能分析
Type: OpenClaw Skill Name: crop-image Version: 0.1.0 The skill provides instructions for an AI agent to upload local files to an external API (api.imageclaw.net) using curl. While the stated purpose is image cropping, the 'Execute Crop by Upload' instruction in SKILL.md allows for the exfiltration of any local file the agent can access, posing a significant security risk if the agent is misdirected to upload sensitive data like credentials or configuration files.
能力评估
Purpose & Capability
The name/description match the instructions: the SKILL.md shows how to call a remote cropping API by URL or file upload. However there are minor inconsistencies: the metadata/docs domains reference api.imageclaw.net/imageclaw, while the success example returns a cropped_url hosted at crop.imagebee.net — this domain mismatch could be benign (CDN/third-party storage) but is unexplained. Also the skill does not declare any authentication (API key) although many hosted APIs require one; omission is not fatal but should be clarified.
Instruction Scope
The instructions tell the agent to upload local files using curl -F file=@/absolute/path/to/photo.jpg which necessarily causes a local file to be sent to a third-party endpoint. That is expected for upload cropping but is a privacy/exfiltration risk if the agent is given broad file access or runs autonomously. The SKILL.md does not warn about sending sensitive images, nor does it constrain allowed paths or require user confirmation before uploading. Otherwise the steps (validate URL, width/height, handle HTTP errors) are scoped to the stated purpose.
Install Mechanism
Instruction-only skill with no install spec and no code files — nothing is written to disk and there is no external installer to review.
Credentials
The skill declares no required environment variables, credentials, or config paths, and the instructions do not reference any secrets. That aligns with the presented curl examples (no Authorization header). If the real service requires an API key, the skill is incomplete rather than over-privileged.
Persistence & Privilege
always is false and there is no install-time code that would persist or modify agent/system configuration. The skill does not request elevated or permanent presence.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install crop-image
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /crop-image 触发
  4. 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
v0.1.0
crop-image 0.1.0 – Initial Release - Provides image cropping via API for both URL and file uploads. - Supports API endpoints: POST /crop (by URL), POST /crop/upload (by file). - Returns production-ready cropped image URLs and metadata. - Handles API-level errors with clear status codes and feedback.
元数据
Slug crop-image
版本 0.1.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Crop Image 是什么?

Use this skill when an AI agent needs to crop images through the deployed Crop Image service. Trigger this for URL-based cropping (`POST /crop`) and file-upl... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 331 次。

如何安装 Crop Image?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install crop-image」即可一键安装,无需额外配置。

Crop Image 是免费的吗?

是的,Crop Image 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Crop Image 支持哪些平台?

Crop Image 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 Crop Image?

由 EricDoe(@effimail)开发并维护,当前版本 v0.1.0。

💬 留言讨论