obsidian-knowledgeize

SKILL效率
效率 未标注权限 许可证不明确
安装这个技能

Convert ordinary Markdown notes into Obsidian knowledge-base notes with frontmatter, tags, bidirectional links, navigation blocks, and knowledge-oriented structure. Use when Codex needs to knowledgeize one or more .md notes, add wiki links to related notes in the same vault, improve note discoverability, or turn linear notes into reusable knowledge cards without reading the entire vault into context.

一句话选型结论

按资料字段整理

obsidian-knowledgeize是效率类的 Agent Skill,安装方式未标注,权限级别未标注,许可证不明确,有可直接复制的安装命令。技能本质是给模型读的说明书,效果取决于模型与客户端;装完先用一个小任务验证,再放进正式流程。

分类效率 安装方式未标注 权限级别未标注权限 许可证许可证不明确
累计安装 未标注 本周 未标注
GitHub 星标 未标注 仓库 未标注
下载量 未标注 技能市场下载数
权限级别 未标注权限 安装前请确认风险范围
许可证 许可证不明确 未说明

安装

未标注
在支持 Skills 的客户端里执行,装完问一句「你有哪些技能」验证
skillhub install user_cff48da7/obsidian-knowledgeize
  1. 1
    确认客户端支持

    确认你的客户端支持 Skills(Claude、Codex、Cursor 等),并升级到较新的版本。

  2. 2
    拿到技能文件

    执行上面的命令安装;也可以把仓库里的 SKILL.md 手动放进客户端的 skills 目录。

  3. 3
    刷新技能列表

    重新打开客户端,或刷新技能列表,让新放进来的技能被加载。

  4. 4
    验证是否生效

    问一句「你有哪些技能」,确认列表里出现它,再拿一个小任务试跑。

技能档案

效率
字段来自技能仓库与公开索引,中文说明由本站标注
作者 / 维护方user_cff48da7仓库 未标注 · 版本 未标注
宿主客户端未标注技能可被这些客户端加载
安装方式未标注安装包 community
权限级别未标注权限安装前请确认风险范围
许可证未说明许可证不明确,商用前请自行确认条款
仓库可用性available技能仓库当前是否可访问
最近核验未标注上游最后核验时间 · 数据更新 1791644811
正文体量5.5 KBSKILL.md 原文大小,越大说明说明书写得越细

典型使用场景

3 类
按分类整理的常见用法,具体以 SKILL.md 为准
整理待办与日程

把零散任务收进清单,排出优先级。

批量处理重复任务

把固定动作做成模板,重复执行。

把流程写成清单

把口头流程整理成可复用的步骤。

选择建议

装之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有可直接复制的安装命令,不用自己翻文档找入口。

先想清楚的地方

  • 没有标注权限级别,实际能做什么要装好后确认。
  • 许可证不明确,商用或二次分发前先到仓库确认条款。
  • 没有做站内核验,内容以作者仓库为准。
  • 技能只是说明书,实际效果取决于模型与客户端,装完先用小任务试一遍。

更适合

  • 需要在效率场景里给模型补一套做法的人。
  • 愿意先读一遍 SKILL.md、确认它会让模型做什么的人。

不太适合

  • 对许可证要求很严、必须明确可商用的人。
  • 只想装完就自动变强、不打算看说明书的人。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
技能权限级别安装方式累计安装宿主客户端安装命令
obsidian-knowledgeize当前 未标注权限 未标注 未标注 未标注 有
Nano PDF 文档 联网访问 ClawHub 市场 424.9 万 OpenClaw 有
自动更新程序 浏览器控制 ClawHub 市场 296.4 万 OpenClaw 有
Gemini 助手 联网访问 ClawHub 市场 254.7 万 OpenClaw 有
word DOCX 文档 联网访问 ClawHub 市场 239.6 万 OpenClaw 有
Excel XLSX 前端 联网访问 ClawHub 市场 227.0 万 OpenClaw 有

SKILL.md 全文

5.5 KB
模型实际读到的说明书原文,来自技能仓库

Obsidian Knowledgeize

Convert a plain Markdown note into an Obsidian-friendly knowledge note while keeping token use under control.

Prefer repository-aware linking over generic rewriting. Improve structure and retrieval value first; rewrite prose only when the user explicitly asks.

Core Rule

Never read the full vault by default.

Use this retrieval order:

  1. Read the target note.
  2. Read only note titles and relative paths across the vault.
  3. Select a small candidate set of potentially related notes.
  4. Read only the top candidate notes whose titles suggest real relevance.
  5. Add links only when the relation is defensible from the target note plus the inspected candidates.

If the vault is large, title-only discovery is the default behavior, not an optimization.

Default Output

When knowledgeizing a note, prefer these upgrades:

  • Add YAML frontmatter with concise tags and aliases when justified.
  • Add a short Knowledge Position or Note Position section near the top.
  • Add Related Notes and Navigation or an equivalent navigation section near the bottom.
  • Add a small set of high-value [[wikilinks]] to related concepts, tasks, models, projects, or hub pages.
  • Add a short Further Split Ideas section when the note is long and contains multiple separable ideas.

Do not force all sections if the note type does not need them.

Workflow

1. Classify the target note

Decide what kind of note it is before linking:

  • concept note
  • summary note
  • project or practice note
  • task note
  • paper note
  • tutorial note
  • index or MOC note

This determines what metadata and links are appropriate.

2. Build a cheap vault index

Discover only titles and paths first. Prefer fast file listing commands such as rg --files -g "*.md" or equivalent.

From paths and filenames, infer:

  • note titles
  • folder themes
  • likely hub pages
  • obvious siblings

Do not open note bodies yet.

3. Choose candidate notes

Select candidates using cheap signals:

  • title token overlap
  • shared entities or acronyms
  • same model or task family
  • same folder or neighboring folder
  • likely parent-child relation such as overview vs detail
  • likely theory-practice relation such as RNN vs RNN Practice

Keep the candidate set small. A good default is 3 to 8 notes.

If a relationship is obvious from title alone, linking without reading the candidate body is acceptable. Example: a note named NLP-KB-Overview.md is clearly a hub page.

4. Read only the candidates that matter

Open candidate bodies only when title evidence is insufficient to justify linking.

Prioritize candidates that may answer one of these questions:

  • Is this really the same concept?
  • Is this note a parent, child, sibling, or hub?
  • Should the link be conceptual, navigational, or comparative?

Stop reading once the link set is good enough. Do not keep exploring for completeness.

5. Edit conservatively

Default to structure-preserving enhancement:

  • keep the original body
  • add metadata and navigation
  • insert a small number of high-value links
  • avoid turning every noun phrase into a wikilink

Prefer precision over density. Too many links reduce note quality.

Link Selection Rules

Add links when at least one of these is true:

  • the linked note is a direct prerequisite
  • the linked note is a direct follow-up
  • the linked note is the same topic at a different granularity
  • the linked note is the theory behind the current practice note
  • the linked note is a hub, map, or overview page
  • the linked note is repeatedly referenced by the target note

Avoid links when:

  • the relation is generic and weak
  • the title match is accidental
  • the concept is too broad to help navigation
  • the vault does not already have a defensible target note

Prefer existing notes over creating many new placeholder notes unless the user asked for expansion.

Link Density Rules

For a normal note, aim for a sparse, useful link set:

  • 1 hub or navigation link
  • 2 to 6 strong related-note links
  • 0 to 5 concept links inside the body, only for central terms

Do not link the same term repeatedly in every paragraph.

Metadata Rules

Keep frontmatter small and stable. Good defaults:

  • tags
  • aliases

Optional fields only when clearly useful:

  • source_type
  • knowledge_type
  • status

Do not invent large metadata schemas unless the vault already uses one.

Body Rewrite Rules

By default:

  • do not rewrite the full note
  • do not aggressively paraphrase
  • do not reorganize every section

You may lightly normalize headings, add short framing blocks, and fix obvious structure issues.

Only do substantial rewriting when the user explicitly asks for summarization, polishing, or refactoring.

Batch Mode

When the user asks to knowledgeize multiple notes:

  • build the title/path index once
  • process notes one by one
  • reuse the same title index
  • read additional note bodies only for each note's top local candidates

Do not preload the whole vault.

References

Read references/linking-strategy.md when you need the detailed candidate-selection and stop-reading heuristics.

常见问题

5 问
答案由本站依据条目字段整理
obsidian-knowledgeize怎么安装?

在支持 Skills 的客户端里执行:skillhub install user_cff48da7/obsidian-knowledgeize。安装方式为未标注。装完刷新技能列表或重开客户端。

obsidian-knowledgeize支持哪些客户端?

资料没有标注宿主客户端,但只要客户端支持 Skills(Claude、Codex、Cursor 等),一般都能加载。

obsidian-knowledgeize权限高吗,安全吗?

权限级别为未标注权限。安装前先看它会读取或写入哪些内容。

obsidian-knowledgeize收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于技能本身和所用的模型或客户端,正式使用前按作者说明确认。

obsidian-knowledgeize用什么许可证?

许可证标注为 未说明(许可证不明确)。商用或二次分发前请到仓库确认条款原文。

热门专题

按场景挑技能
全部专题 →

技能内容以作者仓库为准;分类、权限级别与许可证状态的中文说明为中文整理,最近核验 未标注。