← Back to Skills Marketplace
Image Manager
by
xiaobu2020
· GitHub ↗
· v1.0.0
· MIT-0
176
Downloads
1
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install image-manager
Description
Manage local images by category with original and thumbnail storage, enabling fast tag, date, and keyword search through a JSON index file.
README (SKILL.md)
Image Manager Skill
本地图片管理技能，支持索引、压缩、分类、快速查找。
功能
- 📁 分类存储：按类别（pets/people/food/scenery/receipts/other）自动归档
- 🔍 快速索引：基于 JSON 索引文件，支持按标签、日期、类别查找
- 🗜️ 压缩保存：原图 + 缩略图双存储，查看缩略图不影响原图画质
- 🏷️ 标签系统：每张图片可附加多标签，支持标签搜索
- 📊 摘要浏览：快速浏览某个类别/标签的所有图片
目录结构
media/
├── images/ # 原图（保持完整画质）
│ ├── pets/
│ ├── people/
│ ├── food/
│ ├── scenery/
│ ├── receipts/
│ └── other/
├── thumbnails/ # 缩略图（快速预览）
│ ├── pets/
│ ├── people/
│ └── ...
└── index.json # 全局索引文件
快速使用
保存图片
python scripts/save_image.py \x3Cimage_path> \
--category pets \
--tags "包子,白色,长毛" \
--description "包子坐在地上"
查找图片
# 按标签查找
python scripts/search_image.py --tags "包子"
# 按类别查找
python scripts/search_image.py --category pets
# 按日期查找
python scripts/search_image.py --date 2026-03-19
# 按关键词查找（搜索 description 和 tags）
python scripts/search_image.py --keyword "白色"
浏览摘要
python scripts/list_images.py --category pets
python scripts/list_images.py --tags "包子"
索引格式（index.json）
每条记录包含：
{
"id": "baozi-2026-03-19-001",
"path": "media/images/pets/baozi-2026-03-19-001.jpg",
"thumbnail": "media/thumbnails/pets/baozi-2026-03-19-001.jpg",
"category": "pets",
"tags": ["包子", "白色", "长毛"],
"description": "包子坐在地上",
"saved_at": "2026-03-19T22:33:00+08:00",
"source": "feishu",
"size_bytes": 88688,
"width": 1080,
"height": 1440
}
分类说明
| 类别 | 用途 |
|---|---|
| pets | 宠物照片 |
| people | 人物照片 |
| food | 美食/食物 |
| scenery | 风景/地点 |
| receipts | 小票/账单 |
| other | 未分类 |
设计原则
- 原图不动：保存时不做有损压缩，原图完整保留
- 缩略图预览：生成 300x300 缩略图用于快速浏览
- 索引优先：查找时只读索引文件，不遍历磁盘
- 标签灵活：一张图可有多个标签，标签可随时增删
Usage Guidance
This skill is a local image manager and appears to only read/write under a workspace folder. Before installing: confirm where OPENCLAW_WORKSPACE will point (scripts default to /home/admin/.openclaw/workspace) so you know which directory will get media/index.json and image files; ensure you are comfortable the agent can write there. Note the scripts optionally use Pillow (PIL) to generate thumbnails — if not present they fallback to copying the image as a thumbnail. There are no network calls or secret access. If you need stricter isolation, run the scripts in a controlled directory or inspect index.json/media contents before trusting them.
Capability Analysis
Type: OpenClaw Skill
Name: image-manager
Version: 1.0.0
The image-manager skill bundle is a legitimate tool for local image organization, indexing, and thumbnail generation. The scripts (save_image.py, list_images.py, search_image.py) operate strictly within the defined workspace, use standard libraries for image processing and file management, and contain no evidence of data exfiltration, malicious execution, or prompt injection.
Capability Assessment
Purpose & Capability
Name/description match the provided scripts: they index, store originals and thumbnails, and support tag/date/category search. Requested resources (filesystem under a workspace) are appropriate for the stated purpose.
Instruction Scope
SKILL.md directs running the provided Python scripts with local file paths and expected arguments. The runtime instructions do not ask the agent to read unrelated system files, send data externally, or perform privileged operations.
Install Mechanism
No install spec or external downloads; the skill is instruction + local scripts only. That minimizes installation risk (nothing is fetched or executed from remote URLs).
Credentials
The scripts read OPENCLAW_WORKSPACE (defaulting to /home/admin/.openclaw/workspace) but the skill metadata lists no required env vars. This is reasonable (an optional workspace override) but the env var is not declared in the manifest and the SKILL.md does not document it.
Persistence & Privilege
always is false and the skill does not attempt to modify other skills or system-wide agent settings. It writes only under its workspace media/ directory (index.json, images, thumbnails).
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install image-manager - After installation, invoke the skill by name or use
/image-manager - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Image Manager Skill v1.0.0
- Initial release with local image management capabilities.
- Supports automatic image categorization (pets, people, food, scenery, receipts, other).
- Features dual storage: original image and 300x300 thumbnail for quick preview.
- Provides fast search by tags, category, date, and keywords using a JSON index file.
- Allows multi-tag system and summary browsing by category or tag.
- Index records include metadata such as path, tags, description, date, source, and size.
Metadata
Frequently Asked Questions
What is Image Manager?
Manage local images by category with original and thumbnail storage, enabling fast tag, date, and keyword search through a JSON index file. It is an AI Agent Skill for Claude Code / OpenClaw, with 176 downloads so far.
How do I install Image Manager?
Run "/install image-manager" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Image Manager free?
Yes, Image Manager is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Image Manager support?
Image Manager is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Image Manager?
It is built and maintained by xiaobu2020 (@xiaobu2020); the current version is v1.0.0.
More Skills