← Back to Skills Marketplace
regexl

article-image

by Vincent · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
324
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install article-image
Description
文章配图推荐。根据文章主题、内容关键词,推荐合适的配图来源和搜索关键词,帮助用户找到符合文章意境的图片。当用户提到「配图」「找图」「文章图片」「封面图」「插图」时激活。
README (SKILL.md)

文章配图推荐 Skill

功能概述

根据文章主题和内容,智能推荐配图方案,包括图片来源、搜索关键词、风格建议等。

使用方式

用户可以说:

  • "帮我找几张配图"
  • "这篇文章用什么图片好"
  • "推荐一个封面图"
  • "找一张 xxx 相关的图片"

配图来源

免费图库

网站 特点
Unsplash 高质量自然风景、生活
Pexels 丰富多样、商用免费
Pixabay 涵盖广、中文友好
Unsplash 人文、科技、商业
Pexels 视频也很棒
pixiv 日系插画
Dribbble 设计感强

中文图库

  • 站酷 (ZCOOL)
  • 花瓣网
  • 阿里图标库

付费图库

  • Shutterstock
  • Getty Images
  • 视觉中国

搜索技巧

按主题

文章主题 英文关键词
科技/互联网 technology, coding, digital, AI, innovation
商业/职场 business, office, meeting, professional
生活方式 lifestyle, coffee, nature, travel
教育/学习 education, books, library, student
金融/投资 finance, chart, growth, investment
健康/健身 health, fitness, exercise, yoga

按风格

  • 扁平化 (flat, minimal)
  • 渐变 (gradient, colorful)
  • 插画 (illustration, vector)
  • 照片 (photo, realistic)
  • 抽象 (abstract, pattern)

组合搜索

"keyword1 + keyword2"
"keyword -unwanted"
"site:unsplash.com keyword"

配图原则

1. 相关性

  • 图片与文章主题相关
  • 突出文章核心观点

2. 视觉层次

  • 封面图:大气、吸引眼球
  • 文中图:补充说明、缓解阅读疲劳
  • 题图:简洁、点题

3. 版权合规

  • 确认授权范围
  • 署名要求
  • 商用限制

4. 尺寸适配

  • 封面:1920x1080 或 1200x630
  • 文中:800x450 或 16:9 比例
  • 头像:圆形或正方形

工作流程

  1. 分析文章

    • 提取主题关键词
    • 确定文章风格(正式/轻松/技术)
  2. 搜索图片

    • 直接打开图库网站搜索
    • 不需要询问用户,直接执行
  3. 下载保存

    • 默认保存到:C:\Users\lwp\images\
    • 按规范命名:主题-序号.jpg
    • 无需询问,直接下载 3-5 张备选
  4. 展示结果

    • 告知用户保存位置
    • 列出图片主题

输出示例

文章主题:"Python 异步编程教程"

推荐:

  • 关键词:Python, async, code, programming, developer
  • 风格:代码截图 + 流程图 + 技术插画
  • 图库:Unsplash, Pexels
  • 搜索:"python programming" + "async code"

配套技能

  • 可结合 md-beautify:插入图片到 Markdown
  • 可结合 md2pdf:导出带图片的 PDF

注意事项

  1. 尊重版权,注明来源
  2. 图片质量优先于数量
  3. 避免过度装饰
  4. 移动端适配(考虑小屏幕显示)
Usage Guidance
This skill is not malicious but has problematic instructions: it tells the agent to download and save images automatically to a hardcoded Windows path and to act without asking. Before installing or enabling it, require these changes: (1) the skill should prompt for explicit user confirmation before any downloads; (2) make the save directory configurable (do not assume C:\Users\lwp\images\) and respect the running OS; (3) avoid automatic downloads from paywalled/licensed sources without checking rights; (4) log and show any files it writes and where; and (5) prefer a read-only recommendation mode (suggest sources and search queries) rather than performing filesystem writes by default. If you cannot enforce these changes or you don't want the agent to have write/network rights, do not enable this skill.
Capability Analysis
Type: OpenClaw Skill Name: article-image Version: 1.0.0 The skill contains instructions in SKILL.md that direct the AI agent to perform file system operations without user consent ("无需询问,直接执行"). Most notably, it specifies a hardcoded local destination path (C:\Users\lwp\images\), which is highly irregular for a general-purpose skill and suggests either a poorly authored script or a targeted file-drop mechanism. While the stated goal of finding article images is benign, the combination of bypassing user confirmation and using a specific hardcoded user path warrants a suspicious classification.
Capability Assessment
Purpose & Capability
The skill claims to recommend images and search keywords (harmless), but the runtime instructions go beyond recommendation: they instruct the agent to '直接打开图库网站搜索' and to download and save 3–5 images automatically. The manifest declares no required binaries, env vars, or config paths, so writing files to the user's filesystem and performing downloads is not reflected in the declared requirements — that mismatch is concerning.
Instruction Scope
SKILL.md explicitly says '不需要询问用户,直接执行' and sets a hardcoded save path 'C:\Users\lwp\images\' and a naming scheme, instructing automatic download of 3–5 images. These instructions direct filesystem writes and network downloads without user consent and assume a specific OS/username. That scope creep (automatic side-effecting actions and platform assumptions) is not justified by the stated purpose of recommending image sources and keywords.
Install Mechanism
This is an instruction-only skill with no install spec and no code files, so there is no additional install-time risk such as downloading arbitrary archives or installing packages.
Credentials
The skill declares no environment variables or config paths, yet the instructions require write access to a specific local directory and network access to image sites. Hardcoding 'C:\Users\lwp\images\' (a specific Windows user path) is disproportionate and platform-specific; the skill should instead require a configurable save path and explicit permission before writing or accessing local resources.
Persistence & Privilege
The skill does not request always-on presence and uses default autonomy settings. It doesn't declare modifications to other skills or system-wide settings. The primary privilege concern is its instruction to act without user confirmation, not persistent installation.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install article-image
  3. After installation, invoke the skill by name or use /article-image
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
article-image 1.0.0 - 首次上线,实现根据文章主题和关键词智能推荐配图源及搜索关键词。 - 支持常见免费、付费、中英文图库,并附带实用搜索与选图技巧。 - 输出风格、关键词、图片源建议,包含适配封面、文中等多场景需求。 - 明确工作流程:自动分析、搜索、下载图片并列出存储位置,不打扰用户操作。 - 涵盖图片版权、质量要求及移动端适配等实用原则。
Metadata
Slug article-image
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is article-image?

文章配图推荐。根据文章主题、内容关键词,推荐合适的配图来源和搜索关键词,帮助用户找到符合文章意境的图片。当用户提到「配图」「找图」「文章图片」「封面图」「插图」时激活。 It is an AI Agent Skill for Claude Code / OpenClaw, with 324 downloads so far.

How do I install article-image?

Run "/install article-image" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is article-image free?

Yes, article-image is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does article-image support?

article-image is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created article-image?

It is built and maintained by Vincent (@regexl); the current version is v1.0.0.

💬 Comments