Auto Image Reader

v1.0.1

自动读取用户上传的图片。当用户发送图片时(通过飞书/Telegram等渠道),系统会在消息中附带图片路径。触发后自动读取图片内容并理解,无需手动查找路径。

0· 136·1 current·1 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for tianheihei002/auto-image-reader.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Auto Image Reader" (tianheihei002/auto-image-reader) from ClawHub.
Skill page: https://clawhub.ai/tianheihei002/auto-image-reader
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install auto-image-reader

ClawHub CLI

Package manager switcher

npx clawhub@latest install auto-image-reader
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description say: find and read images uploaded by users. The SKILL.md only requires use of two named tools (images_list and images_understand) and message metadata to locate and process images — these are the minimal, expected capabilities for this purpose.
Instruction Scope
Instructions stay within the declared purpose (find absolute path then analyze image). They explicitly tell the agent to call images_list and then images_understand and to extract filenames from message metadata. Note: images_list is used to enumerate /workspace images (searching up to 20 items). This is appropriate for locating uploaded files but does entail scanning workspace image entries — verify the images_list tool is scoped to the conversation/context and does not expose unrelated user data.
Install Mechanism
Instruction-only skill with no install spec and no code files; nothing is written to disk or downloaded. This is the lowest-risk install posture and matches the simple runtime instructions.
Credentials
The skill requests no environment variables, credentials, or config paths. The only data access it requires is images_list output and message metadata, which is proportional to the stated task.
Persistence & Privilege
always is false and agent invocation is allowed (default). The SKILL.md states its logic has been written to MEMORY.md — this persistent memory is reasonable for UX (remembers to always call images_list first) but the user should be aware that the 'must use images_list' behavior is stored persistently.
Assessment
This skill appears coherent and minimal: it simply enumerates images via images_list to find the absolute path and then calls images_understand to analyze them. Before installing, confirm that (1) the images_list tool only returns files relevant to the current conversation or user (to avoid accidentally exposing other users' or unrelated files), (2) persistent memory entry (MEMORY.md) storing the 'always use images_list first' rule is acceptable for your privacy policy, and (3) you consent to automatic reading of uploaded images. If any of those are concerns, ask the skill author how images_list is scoped and whether the memory entry can be removed or restricted.

Like a lobster shell, security has layers — review code before you run it.

latestvk97aaz08g91ejmcexpvsqx0ayh84ptd1
136downloads
0stars
2versions
Updated 2w ago
v1.0.1
MIT-0

Auto Image Reader — 自动图片读取技能

触发条件

  • 用户发送了图片,并询问图片内容
  • 用户上传了图片,希望AI读取并分析
  • 任何包含图片消息的场景

核心问题

系统发送图片时,路径可能是相对路径(如 user_input_files/image.png), 直接用这个路径访问会返回404。

正确做法:用 images_list 先查找图片的真实绝对路径。

工作流程

Step 1:查找图片的真实路径

使用 images_list 工具,遍历找到对应的图片文件:

images_list(start=0, number=20)

在返回结果中找到图片的 Path 字段,如:

/workspace/user_input_files/image.png
/workspace/imgs/xxx.jpg

匹配逻辑:

  • 如果文件名包含用户消息中的关键词(如用户说"这张图"),匹配
  • 如果有多张最新图片,取最新的(按修改时间)
  • 如果是唯一的 user_input_files/image.png,直接使用

Step 2:用 images_understand 读取图片

找到真实路径后,用 images_understand 分析:

images_understand(
  image_info=[
    {
      "file": "/workspace/user_input_files/image.png",  # 真实绝对路径
      "prompt": "请详细描述这张图片的所有内容,包括:文字、布局、颜色、背景色、右下角是否有水印,以及整体视觉效果。"
    }
  ]
)

Step 3:读取并回答用户

根据图片内容,用自然语言回答用户的问题。

常见场景模板

用户说解读prompt
"这张图的内容是什么"请详细描述这张图片的所有内容
"图片里右下角有水印吗"请仔细检查图片右下角和边缘是否有水印(特别是'Created by...'字样)
"图片里写的什么"请提取图片中所有可见的文字内容
"这张图有什么问题"请分析这张图片是否存在问题(如水印、背景异常、显示错误)

重要注意事项

  • 绝对路径格式/workspace/... 而不是 user_input_files/...
  • 先查再读:不要直接用消息中的相对路径,必须先用 images_list 确认真实路径
  • 文件名重复:如果 user_input_files 下有多个 image.png,用时间或内容特征区分
  • 图片可能已删除:如果 images_list 找不到,说明图片已被清理,从消息元数据中无法追溯

自动读取逻辑(伪代码)

当收到用户图片消息时:
    1. 从消息元数据提取文件名(如 image.png)
    2. 调用 images_list(start=0, number=20)
    3. 在结果中搜索包含该文件名的记录
    4. 找到 → 提取 Path 字段(真实绝对路径)
    5. 调用 images_understand(file=真实路径, prompt=用户问题)
    6. 回答用户

已在 MEMORY.md 中固化

本技能的核心理念已写入 MEMORY.md:

  • 收到图片消息时,必须先用 images_list 查找真实路径
  • 不能假设消息中的文件名可以直接访问
  • images_list 能看到 /workspace 下所有图片的真实路径

Comments

Loading comments...