Token导航 LogoToken导航TokenDH.com
研究检索需要联网clawhub未标认证来源可访问clear审计提醒

pkmPKM 搜索

Agent Skill

pkm 用于查找、检索和筛选相关信息,适合在 OpenClaw 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。

总安装

93,139

周安装

3,960

GitHub Stars

4

下载量

32,630
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

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:pkm(PKM 搜索)
来源仓库:https://github.com/ivangdavila/pkm
安装命令:
openclaw skills install pkm
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install pkm

简介

pkm 用于查找、检索和筛选个人知识管理相关内容,适合在 OpenClaw 中将零散信息转化为结构化笔记。

  • 它支持分类存储和关联检索功能。
  • 使用方式包括输入原始文本和分类标签,由 Agent 输出笔记模板。
  • 安装命令为 openclaw skills install pkm,需确认本地存储权限。
  • 建议定期整理分类体系,保持知识库可用性。

SKILL.md

name
Personal Knowledge Base
description
Help users build a personal knowledge base by organizing whatever they send into structured notes.
metadata
{"clawdbot":{"emoji":"🧠","os":["linux","darwin","win32"]}}

Core Behavior

  • User sends anything: link, idea, quote, snippet, question, rambling thought
  • Capture first, organize second — never lose input while deciding where it goes
  • Create ~/kb/ as the workspace — flat folder of Markdown files initially
  • Inbox pattern: inbox.md for quick capture, process later into proper notes

When User Sends Content

  • Link → fetch title and summary, save with source URL and capture date
  • Idea/thought → save as atomic note with descriptive filename
  • Quote → save with attribution, link to source if available
  • Question → save as note, mark for future research
  • Long rambling → extract key points, save as separate atomic notes

File Naming Convention

  • Lowercase with hyphens: how-to-negotiate-salary.md
  • Descriptive over date-based — findable by topic, not when captured
  • No rigid hierarchy initially — flat folder with good names beats complex structure
  • Date prefix optional for journals: 2024-01-15-weekly-review.md

Note Structure

  • Title as H1 — matches filename concept
  • Tags at top or bottom — #productivity #career for filtering
  • Source/reference if applicable — where it came from
  • Related notes section — manual links build knowledge graph
  • Keep notes atomic — one concept per note, link between them

Inbox Processing

  • Periodically ask: "Want to process your inbox?"
  • For each item: create proper note, add tags, link to related notes
  • Delete from inbox once processed — inbox should trend toward empty
  • Don't force immediate organization — capture friction kills usage

When To Add Structure

  • 20+ notes: suggest consistent tagging system
  • 50+ notes: suggest index.md or MOC (Map of Content) for key topics
  • 100+ notes: suggest folder structure by domain if patterns emerge
  • Only add structure when navigation becomes painful

Tagging Strategy

  • Start with 5-10 broad tags maximum — too many defeats purpose
  • Tags are for retrieval, not categorization — "when would I search for this?"
  • Multi-tag allowed — note about salary negotiation: #career #communication
  • Review and consolidate tags periodically — synonyms fragment knowledge

Linking Between Notes

  • [[wiki-style]] links when supported, otherwise relative Markdown links
  • Link liberally — connections are the value of knowledge base
  • Backlinks show where note is referenced — surface hidden connections
  • Don't force links — some notes are standalone

What User Might Send

  • "Just learned that..." → atomic note with insight
  • "Interesting article: [URL]" → fetch, summarize, save with source
  • "Reminder: X" → capture with context, might become action or reference
  • "I keep forgetting how to..." → create or update how-to note
  • Random thought → inbox immediately, process later

Searching and Retrieval

  • Full-text search with grep or specialized tool — must be fast
  • Search by tag: find all notes with specific tag
  • Recent notes list — often want "that thing I saved last week"
  • Offer to search when user asks a question — might already have the answer

Progressive Enhancement

  • Week 1: inbox.md only, dump everything
  • Week 2: process inbox into atomic notes with tags
  • Week 3: start linking related notes
  • Month 2: create index/MOC for main topics
  • Month 3: folder structure if needed

What NOT To Suggest Early

  • Complex folder hierarchies — flat with good names first
  • Database or app — Markdown files work until they don't
  • Daily notes system — unless they specifically want journaling
  • Templates — organic structure emerges, then standardize

Sync and Backup

  • Cloud folder (Dropbox/iCloud) for multi-device access
  • Git repo for version history — see how thinking evolved
  • Plain Markdown ensures portability — not locked to any tool

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

78.52%
按下载量换算25,621

安全审计

VirusTotal

可疑

ClawScan

通过

Static analysis

未展示

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills