Token导航 LogoToken导航

Agent Skill · 研究检索

iib(国际银行)

iib 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

iib 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为操作浏览器;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 62,244。

总安装 339周安装 14GitHub Stars 1,269下载 111核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:iib(国际银行)
来源仓库:https://github.com/zanllp/infinite-image-browsing
仓库路径:skills/iib
安装命令:
npx skills add https://github.com/zanllp/infinite-image-browsing --skill iib
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/zanllp/infinite-image-browsing --skill iib

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

iib 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。适用于需要根据关键词或任务场景从来源线索中提取信息的场景。通过关键词、任务场景或来源线索进行信息检索与筛选。安装前建议确认权限范围、维护状态及是否触发联网或命令执行。

可结合来源仓库和原始 README 进一步核验具体用法。

来源仓库内路径:skills/iib

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/zanllp/infinite-image-browsing --skill iib;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要操作浏览器,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

iib 怎么安装?

在终端执行 npx skills add https://github.com/zanllp/infinite-image-browsing --skill iib。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

iib 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

iib 需要什么权限?

权限级别是操作浏览器。该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

iib 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

iib 用的人多吗?

公开数据是累计安装 339 次,近一周安装 14 次,下载 111 次,GitHub 1,269 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

iib 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 来源仓库有 1,269 星,维护活跃度相对可查。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
iib

国际银行 · 本页

3391,269操作浏览器未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

IIB (Infinite Image Browsing)

IIB is an image/video browsing and management tool that parses metadata from AI generation tools (Stable Diffusion, ComfyUI, etc.).


Quick Links to Detailed Guides

TopicDescription
API ReferenceComplete API endpoint documentation
Search StrategiesMulti-word, regex, tag combination searches
Agent PatternsCommon workflows and decision trees

Before You Start

IMPORTANT: Always do these two things first:

  1. Ask the user for the port if they started the service themselves (common ports: <port> standalone, 7860 SD WebUI extension)
  2. Test connectivity with a hello request before any other operation
curl --noproxy "*" -s http://127.0.0.1:<PORT>/infinite_image_browsing/hello
# Returns: "hello" if service is running

Note: Use --noproxy "*" to bypass proxy for localhost connections.

If service is not running, start it:

cd /path/to/sd-webui-infinite-image-browsing
python app.py --port <port>

To run as a background daemon:

nohup python app.py --port <port> > iib.log 2>&1 &

Quick Reference

TaskMethodEndpoint
Search imagesPOST/db/search_by_substr
Search by tagsPOST/db/match_images_by_tags
Random imagesGET/db/random_images
List folderGET/files?folder_path=...
Move filesPOST/move_files
Copy filesPOST/copy_files
Tag imagesPOST/db/batch_update_image_tag
Get image metadataGET/image_geninfo?path=...
Add library pathPOST/db/extra_paths
Remove library pathDELETE/db/extra_paths

Base URL: http://127.0.0.1:<port>/infinite_image_browsing

Core Operations

Search Images

Basic Keyword Search

Search by keyword in file path or generation parameters:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/search_by_substr \
  -H "Content-Type: application/json" \
  -d '{"surstr": "landscape", "size": 50}'
Advanced Search

For complex searches (multi-word, regex patterns, tag combinations), see Search Strategies.

Quick tips:

  • Multi-word searches? Use regex OR: (word1|word2)
  • Need tag combinations? Use /db/match_images_by_tags with AND/OR/NOT logic
  • See Search Strategies for detailed strategies
Limit to Specific Folders
curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/search_by_substr \
  -H "Content-Type: application/json" \
  -d '{"surstr": "portrait", "folder_paths": ["/path/to/folder"], "size": 50}'

Response:

{
  "files": [{"fullpath": "/path/to/image.png", "name": "image.png", "size": "1.2 MB", ...}],
  "cursor": {"has_next": true, "next": "cursor_string"}
}

Tag Management

Get all tags:

curl http://127.0.0.1:<port>/infinite_image_browsing/db/basic_info
# Response includes: {"tags": [{"id": 1, "name": "favorites", "type": "custom"}, ...]}

Create a tag:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/add_custom_tag \
  -H "Content-Type: application/json" \
  -d '{"tag_name": "favorites"}'

Tag multiple images:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/batch_update_image_tag \
  -H "Content-Type: application/json" \
  -d '{"img_paths": ["/path/to/img1.png", "/path/to/img2.png"], "action": "add", "tag_id": 1}'

Search by tags (AND/OR/NOT logic):

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/match_images_by_tags \
  -H "Content-Type: application/json" \
  -d '{"and_tags": [1], "or_tags": [], "not_tags": [2], "cursor": "", "size": 50}'

File Operations

List folder contents:

curl "http://127.0.0.1:<port>/infinite_image_browsing/files?folder_path=/path/to/images"

Move files:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/move_files \
  -H "Content-Type: application/json" \
  -d '{"file_paths": ["/path/to/img1.png"], "dest": "/new/folder", "create_dest_folder": true}'

Copy files:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/copy_files \
  -H "Content-Type: application/json" \
  -d '{"file_paths": ["/path/to/img1.png"], "dest": "/backup/folder", "create_dest_folder": true}'

Delete files:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/delete_files \
  -H "Content-Type: application/json" \
  -d '{"file_paths": ["/path/to/unwanted.png"]}'

Image Metadata

Get generation parameters (prompt, seed, model, etc.):

curl "http://127.0.0.1:<port>/infinite_image_browsing/image_geninfo?path=/path/to/image.png"

Batch get metadata:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/image_geninfo_batch \
  -H "Content-Type: application/json" \
  -d '{"paths": ["/path/to/img1.png", "/path/to/img2.png"]}'

Library Management

List registered paths:

curl http://127.0.0.1:<port>/infinite_image_browsing/db/extra_paths

Add a folder to library:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/extra_paths \
  -H "Content-Type: application/json" \
  -d '{"path": "/new/image/folder", "types": ["scanned"]}'

Remove a folder (also cleans up orphaned image records):

curl -X DELETE http://127.0.0.1:<port>/infinite_image_browsing/db/extra_paths \
  -H "Content-Type: application/json" \
  -d '{"path": "/old/folder", "types": ["scanned"]}'

Rebuild index after adding paths:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/rebuild_index

Path types:

  • scanned: Indexed, appears in search results
  • scanned-fixed: Like scanned, but pinned in UI
  • walk: Browse only, not indexed

AI Features

Smart File Organization

Automatically organize images into themed folders:

# Start organization job
curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/organize_files_start \
  -H "Content-Type: application/json" \
  -d '{
    "folder_paths": ["/messy/folder"],
    "dest_folder": "/organized/folder",
    "threshold": 0.85,
    "lang": "en",
    "action": "move"
  }'
# Returns: {"job_id": "uuid"}

# Check status
curl "http://127.0.0.1:<port>/infinite_image_browsing/db/organize_files_status?job_id=<job_id>"

# Confirm and execute
curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/organize_files_confirm \
  -H "Content-Type: application/json" \
  -d '{"job_id": "<job_id>"}'

Image Clustering

Analyze images and group by semantic similarity:

curl -X POST http://127.0.0.1:<port>/infinite_image_browsing/db/cluster_iib_output_job_start \
  -H "Content-Type: application/json" \
  -d '{
    "folder_paths": ["/path/to/images"],
    "threshold": 0.85,
    "min_cluster_size": 3,
    "lang": "en",
    "recursive": true
  }'

Reference

See references/api-reference.md for complete API documentation.

Viewing Images

When helping users view images, always provide BOTH:

  1. A link to view the first/selected image
  2. A link to the search results page showing all matching images

Image Search Results

When a user searches for images (by keyword, tag, etc.):

  1. Provide a summary: Show how many images were found
  2. First image link: Use the quick view URL for the first result
  3. Search results page link: Use the fuzzy-search or tag-search pane URL

Example response format:

Found 150+ images matching "sunset":

[View First Image](http://127.0.0.1:<port>/infinite_image_browsing?action=view&path=/first/image.png)

[View All Results in IIB](http://127.0.0.1:<port>/infinite_image_browsing?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22sunset%22%7D)

| # | Name | Path | Size |
|---|------|------|------|
| 1 | image1.png | /folder | 1.2 MB |
| 2 | image2.png | /folder | 1.1 MB |
...

Note: The search results page link includes the search term in the URL, so it will automatically display the matching results when opened.

Building Shareable URLs

When generating URLs for users to open in their browser, you need to properly encode the props parameter (JSON URL-encoded).

URL Format:

http://127.0.0.1:<port>/infinite_image_browsing?action=pane&type=<pane_type>&props=<url_encoded_json>

Encoding the props: The props parameter must be URL-encoded JSON. Use this encoding method:

JSON PropsURL EncodedNotes
{"substr":"sunset"}%7B%22substr%22%3A%22sunset%22%7DStandard encoding
`{"substr":"(词1\词2)","isRegex":true}`%7B%22substr%22%3A%22(%E8%AF%8D1%7C%E8%AF%8D2)%22%2C%22isRegex%22%3Atrue%7DChinese chars encoded
`{"substr":"a\b","isRegex":true}`%7B%22substr%22%3A%22(a%7Cb)%22%2C%22isRegex%22%3Atrue%7DPipe `\ encoded as %7C`

Common URL patterns:

Use CaseURL Pattern
Simple keyword search?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22keyword%22%7D
Regex OR search?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22(word1%7Cword2)%22%2C%22isRegex%22%3Atrue%7D
Search in folder?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22keyword%22%2C%22searchScope%22%3A%22%2Fpath%22%7D
View single image?action=view&path=/path/to/image.png

Tip: When generating URLs with non-ASCII characters (Chinese, Japanese, etc.), use proper URL encoding (UTF-8 percent encoding).

Example: Multi-word search with regex (Chinese)

# Search for "太空电梯" OR "三体"
# JSON: {"substr":"(太空电梯|三体)","isRegex":true}
# Encoded: %7B%22substr%22%3A%22(%E5%A4%AA%E7%A9%BA%E7%94%B5%E6%A2%AF%7C%E4%B8%89%E4%BD%93)%22%2C%22isRegex%22%3Atrue%7D
?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22(%E5%A4%AA%E7%A9%BA%E7%94%B5%E6%A2%AF%7C%E4%B8%89%E4%BD%93)%22%2C%22isRegex%22%3Atrue%7D

Available View Options

1. Quick View (Single Image)
http://127.0.0.1:<port>/infinite_image_browsing?action=view&path=/path/to/image.png

Opens IIB directly to the image in fullscreen preview.

2. Keyword Search Page
http://127.0.0.1:<port>/infinite_image_browsing?action=pane&type=fuzzy-search

Opens the keyword search page where users can search.

With pre-filled search term:

http://127.0.0.1:<port>/infinite_image_browsing?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22sunset%22%7D

Automatically searches for "sunset" when opened.

Available props for fuzzy-search (URL → Component mapping):

URL PropComponent PropTypeDescription
substrinitialSubstrstringSearch keyword or regex pattern
isRegexinitialIsRegexbooleanUse regex mode (default: false)
searchScopesearchScopestringFolder path to limit search
pathOnlyinitialPathOnlybooleanSearch only file paths (default: false)
mediaTypeinitialMediaTypestringFilter by media type: "all", "image", "video"
autoSearchautoSearchbooleanAuto-trigger search (default: true when substr is set)

Note: Use the URL prop names (left column) when constructing URLs. The component prop names (middle column) are for internal use.

Examples:

# Regex search for multiple keywords (OR logic)
?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22(sunset%7Csunrise%7Cdawn)%22%2C%22isRegex%22%3Atrue%7D

# Search in specific folder
?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22portrait%22%2C%22searchScope%22%3A%22%2Foutputs%22%7D

# Search images only
?action=pane&type=fuzzy-search&props=%7B%22substr%22%3A%22landscape%22%2C%22mediaType%22%3A%22image%22%7D
3. Tag Search Page
http://127.0.0.1:<port>/infinite_image_browsing?action=pane&type=tag-search

Opens the tag search page for filtering by tags.

4. Tag Search Results (Pre-filtered)
http://127.0.0.1:<port>/infinite_image_browsing?action=pane&type=tag-search-matched-image-grid&props=%7B%22selectedTagIds%22%3A%7B%22and_tags%22%3A%5B1%2C2%5D%7D%7D%7D

Shows images matching specific tags (replace tag IDs with actual values).

5. Image Comparison
http://127.0.0.1:<port>/infinite_image_browsing?action=pane&type=img-sli&props=%7B%22left%22%3A%7B%22fullpath%22%3A%22%2Fa.png%22%7D%2C%22right%22%3A%7B%22fullpath%22%3A%22%2Fb.png%22%7D%7D

Compare two images side by side.

Note: The props parameter must be URL-encoded JSON.

平台分布

Codex

33.23%
按下载量换算37

Claude

30%
按下载量换算33

Cursor

19.93%
按下载量换算22

Gemini CLI

10.04%
按下载量换算11

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills