Token导航 LogoToken导航

Agent Skill · 研究检索

obsidian-organize(Obsidian organize 搜索)

obsidian-organize 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。

obsidian-organize 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,权限为只读;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 65,351。

总安装 282周安装 12GitHub Stars 公开资料未说明下载 99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:obsidian-organize(Obsidian organize 搜索)
来源仓库:https://github.com/huangmingwang/obsidian-organize
仓库路径:skills/obsidian-organize
安装命令:
npx skills add https://github.com/huangmingwang/obsidian-organize --skill obsidian-organize
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/huangmingwang/obsidian-organize --skill obsidian-organize

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。

用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词快速定位候选结果。通过 npx 命令从指定 GitHub 仓库安装,提供组织整理功能。需确认权限范围和维护状态,避免触发不必要的联网或文件操作。

obsidian-organize 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/obsidian-organize

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/huangmingwang/obsidian-organize --skill obsidi…;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

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

obsidian-organize 怎么安装?

在终端执行 npx skills add https://github.com/huangmingwang/obsidian-organize --skill obsidian-organiz…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

obsidian-organize 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。

obsidian-organize 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

obsidian-organize 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。

obsidian-organize 用的人多吗?

公开数据是累计安装 282 次,近一周安装 12 次,下载 99 次。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。

obsidian-organize 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。
  • 标注为只读权限,试用阶段不容易误改本地文件。

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。
  • 只想先试试效果、不希望它改动本地文件的人。

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
obsidian-organize

Obsidian organize 搜索 · 本页

282—只读未明确通过

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

Obsidian Note Auto-Organizer

Automatically organize Obsidian notes using PARA + MOC + lightweight Zettelkasten methodology.

Two Modes

  • Mode A: Organize existing notes — User points to a note (or batch), you categorize, simplify, link, and index it.
  • Mode B: Generate new notes — User gives a topic, you create a well-structured note with full metadata, links, and MOC entry.

Vault Path & Configuration

The vault root is the current working directory.

If .obsidian-organize.yml EXISTS:

  • Read it and use its rules for all categorization, MOC creation, and folder decisions.

If .obsidian-organize.yml does NOT exist — run Auto-Init:

  1. Scan the vault: Use Glob to find all .md files. Read the first 20 lines of each note (title, headings, key terms).
  2. Cluster by topic: Analyze all notes and identify 5-15 natural topic clusters based on content similarity, shared terminology, and domain patterns.
  3. Generate config: For each cluster, derive:

- An Area name (concise, 2-5 characters in the vault's primary language) - 3-8 representative keywords for auto-categorization - 2-5 suggested MOC sub-groups

  1. Also detect:

- Resources sub-folders (PDFs, images, non-technical content) - The vault's primary language (for language field)

  1. Present the generated config to the user for review:

- Show the proposed Areas, keywords, and sub-groups in a readable format - Ask: "这个分类结构可以吗?有需要调整的吗?"

  1. After user confirms: Write .obsidian-organize.yml to vault root.
  2. Then proceed with the user's original request (organize/generate note).

This ensures fully automatic onboarding — new users just run the skill and everything is set up.


Mode A: Organize Existing Notes

Accept a single file path, glob pattern (e.g., **/Java*.md), or topic keyword (e.g., "organize all Java notes") for batch processing.

For each note, execute these steps in order:

Step 1: Read Config & Analyze

  1. Read .obsidian-organize.yml from vault root to load categorization rules.
  2. Read the target note. Determine its primary topic and category by matching content against the areas keywords in config.
  3. Read references/folder-structure.md for the generic PARA structure and decision logic.

Step 2: Generate/Complete Frontmatter

Read references/frontmatter-spec.md for the schema.

  • If no frontmatter exists: generate all fields from content analysis.
  • If frontmatter exists: preserve existing fields, fill missing ones, update updated date.

Step 3: Ask Whether to Simplify Content

If the user chooses 1:

  • Skip Step 3a and Step 4 entirely.
  • Proceed directly to Step 5 (Backup) → Step 6 (Move) → Step 7 (Links) → Step 8 (MOC).
  • Do NOT modify the note's body content at all.

If the user chooses 2:

  • Proceed with content simplification below.

Step 3a: Simplify Content (only if user opted in)

Read references/content-restructuring.md for rules.

  • Remove redundant content, filler phrases, copy-pasted raw text.
  • Preserve core concepts, minimal code examples, personal insights, structured content.
  • Reformat to the standard note template.

Step 4: Evaluate Split (only if user opted in to simplification)

If the note exceeds 300 lines after simplification:

  • Identify logical split points (distinct sub-topics at H2 level).
  • Present numbered split suggestions with proposed new titles.
  • Only split after user confirmation.
  • Each sub-note gets its own frontmatter and is processed through Steps 5-8.

Step 5: Backup Original

Copy the original file to Archives/原始版本/. If a file with the same name already exists, append a timestamp suffix (e.g., Java_20260314.md).

Step 6: Move to Correct Folder

Move the note to the PARA folder matching its category frontmatter field. Create the target folder if it doesn't exist. If a file with the same name exists at the destination, ask the user before overwriting.

Step 7: Insert Bidirectional Links

Read references/moc-and-linking.md for rules.

  • Use Grep to efficiently find related notes by searching for title/keyword matches.
  • Add a ## 相关笔记 section at the end of the note.
  • Update related notes to link back (bidirectional).

Step 8: Update MOC

Read references/moc-and-linking.md for MOC format.

  • Find or create the corresponding MOC in MOCs/.
  • Add the note to the appropriate sub-group.

Mode B: Generate New Note

Step 1: Understand the Topic

Ask clarifying questions if the topic is ambiguous. Determine the target category.

Step 2: Generate Content

Create the note following the simplified format standard from references/content-restructuring.md.

  • Write concise, clear content in Chinese.
  • Follow the standard template: one-line summary, bullet points, details, code examples (if relevant).

Step 3: Complete Frontmatter

Auto-fill all frontmatter fields per references/frontmatter-spec.md. Set status: active (since it's already in clean format).

Step 4: Save to Correct Folder

Save directly to the PARA folder (skip Inbox). Read references/folder-structure.md for categorization.

Step 5: Insert Bidirectional Links

Same as Mode A Step 7.

Step 6: Update MOC

Same as Mode A Step 8.


Batch Processing

When processing multiple notes:

  • Process one note at a time, showing progress (e.g., "Processing 3/15: Java线程池.md").
  • For content simplification, show diff for each note and wait for confirmation.
  • Offer an "approve all remaining" shortcut if the user is satisfied with the quality after reviewing a few.

Safety Rules

  • ALWAYS backup before modifying any note.
  • ALWAYS show diff before simplifying content — never auto-write.
  • NEVER delete notes without explicit user confirmation.
  • NEVER overwrite an existing note at the destination without asking.
  • Create target folders if they don't exist.
  • If categorization is ambiguous, ask the user rather than guessing.

Language

The vault is primarily in Chinese. Generate all content (frontmatter summary, tags, note content) in Chinese. Use Chinese folder names as defined in the taxonomy.

平台分布

Codex

34.58%
按下载量换算34

Claude

29.31%
按下载量换算29

Cursor

19.03%
按下载量换算19

Gemini CLI

8.82%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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