← 返回 Skills 市场
1knownothing

LLM-Wiki-skills

作者 1knownothing · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
186
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install llm-wiki-skills
功能描述
LLM-powered personal wiki knowledge base system. Use this when user wants to build and maintain a persistent wiki using LLMs following the LLM Wiki pattern (...
使用说明 (SKILL.md)

Wiki Knowledge Base - Master Skill

This is a comprehensive system for building and maintaining a personal wiki knowledge base using LLMs.

What This System Does

The LLM Wiki pattern differs from traditional RAG:

  • RAG: LLM retrieves relevant chunks at query time, rediscovering knowledge from scratch
  • Wiki: LLM incrementally builds and maintains a persistent wiki with cross-references already in place

Core Operations

This system provides 5 skills for wiki management:

1. wiki-init

Initialize a new wiki with:

  • Two-layer architecture (wiki content, schema)
  • Directory structure for entities, concepts, sources, syntheses
  • Initial index.md and log.md
  • CLAUDE.md schema with conventions

Note: Source files should be stored externally (e.g., in project's raw/ folder) - the wiki stores summaries and analysis, not copies of source documents.

Trigger: User wants to set up a new wiki

2. wiki-ingest

Process new sources:

  • Read and analyze source documents
  • Create source summary pages
  • Update entity and concept pages
  • Flag contradictions with existing claims
  • Update index.md and log.md

Trigger: User adds a new source or wants to process a document

3. wiki-query

Answer questions against the wiki:

  • Search index.md first
  • Read relevant pages
  • Synthesize answers with citations
  • Support multiple output formats (markdown, tables, slides)
  • File valuable answers back to wiki

Trigger: User asks questions about wiki content

4. wiki-lint

Health check the wiki:

  • Find contradictions between pages
  • Identify stale claims superseded by new sources
  • Find orphan pages with no inbound links
  • Detect missing pages for important concepts
  • Identify missing cross-references
  • Suggest data gaps to fill

Trigger: User wants to maintain or clean up the wiki

5. wiki-maintain

Manage wiki schema:

  • Update CLAUDE.md conventions
  • Modify page templates
  • Refine workflows
  • Add new page types

Trigger: User wants to update wiki rules or structure

Usage

Select the appropriate sub-skill based on the operation:

  • New wiki? → wiki-init
  • Adding content? → wiki-ingest
  • Asking questions? → wiki-query
  • Checking quality? → wiki-lint
  • Updating rules? → wiki-maintain

Directory Structure

When initialized, the wiki follows this structure:

wiki/
├── entities/              # Entity pages
├── concepts/              # Concept pages
├── sources/              # Source summaries
├── syntheses/            # Cross-source syntheses
├── comparisons/          # Comparison analyses
├── overviews/            # Overview pages
├── schema/CLAUDE.md      # Wiki conventions
├── index.md              # Content catalog
└── log.md                # Operation log

# Source files should be stored in the project root's raw/ folder (or similar external location)

Key Concepts

  • index.md: Content-oriented catalog, updated on every ingest
  • log.md: Chronological record with format ## [YYYY-MM-DD] operation | Title
  • CLAUDE.md: Schema defining conventions and workflows
  • Cross-references: Always use [[wiki/page]] link format
安全使用建议
This skill appears internally consistent for building and maintaining a local LLM-driven wiki, but take these practical precautions before installing: - Confirm the exact directory the skill will use. The SKILL.md instructs the agent to ask you for the wiki location; answer that explicitly and verify the agent will operate only in that path (e.g., a dedicated wiki/ and raw/ folder). - Be aware the agent will read files in the specified raw/ folder and will write files under the wiki/ tree (index.md, log.md, and many content files). Do not point it at folders containing sensitive or unrelated data. - If you plan to integrate with Obsidian or other CLIs, ensure those tools are installed and that you understand what the skill will do with them; the skill references Obsidian as an optional integration but does not require it. - Review and, if desired, edit the SKILL.md to tighten allowed paths or add safeguards (e.g., denylist of directories, confirmation prompts before bulk edits) before granting the agent filesystem access. - Because the skill can be invoked autonomously by agents (platform default), consider whether you want automatic runs; if not, restrict invocation in your agent settings. If you want higher assurance, ask the skill author for a manifest that explicitly limits filesystem scope or provides example runs in a sandboxed directory; otherwise this skill can be used safely with the above precautions.
功能分析
Type: OpenClaw Skill Name: llm-wiki-skills Version: 1.0.0 The skill bundle implements a structured 'LLM Wiki' system for persistent knowledge management, allowing an agent to initialize, ingest, query, and maintain a markdown-based knowledge base. The instructions in the SKILL.md files are well-defined and align with the stated purpose of organizing research and notes (specifically focused on AI interpretability in the provided examples). No evidence of data exfiltration, malicious execution, or harmful prompt injection was found; the system relies on standard local file operations and provides clear workflows for the agent to follow.
能力评估
ℹ Purpose & Capability
The skill declares no required binaries or credentials and is instruction-only, which is broadly consistent with a filesystem-based wiki manager. Minor mismatch: the README and usage notes reference Obsidian, Obsidian CLI, and a web-clipper integration but the package does not declare any required binaries or integrations; these are optional integration suggestions rather than enforced dependencies. Example files include absolute Windows paths (D:\obsidian-notebook\raw\...)—these are example artifacts and the SKILL.md correctly instructs the agent to ask the user for the wiki location before creating files.
ℹ Instruction Scope
Runtime instructions explicitly tell the agent to read index.md and other wiki pages, read source files from an external raw/ folder, synthesize answers, and write or update pages (sources/, concepts/, entities/, syntheses/, index.md, log.md). This behavior is coherent for the stated purpose. Important practical note: because the skill uses Read/Write/Glob/Grep/Bash capabilities, it will operate on files the agent can access—confirming the exact target directories with the user is critical to avoid unintended reads of unrelated files. The instructions do not direct data to external network endpoints or require secrets.
✓ Install Mechanism
No install spec and no code files — instruction-only skills are the lowest-risk class and nothing is downloaded or executed. There is no installer pulling remote artifacts, so install risk is minimal.
✓ Credentials
The skill declares no required environment variables, credentials, or config paths. Its file reads/writes (wiki and external raw/ sources) are proportionate to a wiki ingest/query/maintenance tool.
✓ Persistence & Privilege
The skill does write files (creates/updates wiki pages, index.md, log.md) but 'always' is false and it does not request elevated or cross-skill configuration changes. Autonomous invocation is allowed (platform default) but is not combined with any unusual privileges.
如何使用
  1. 确保已安装 OpenClaw（本地或 Docker 部署）
  2. 在对话框中输入安装命令：/install llm-wiki-skills
  3. 安装完成后，直接呼叫该 Skill 的名称或使用 /llm-wiki-skills 触发
  4. 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
基于karpathy的LLM-wiki方法构建的skills。用于指导agent调用obsidian CLI工具构建个人知识库。Skills built on Karpathy’s LLM-wiki methodology, designed to guide the agent in invoking the Obsidian CLI to build a personal knowledge base.
元数据
Slug llm-wiki-skills
版本 1.0.0
许可证 MIT-0
累计安装 1
当前安装数 1
历史版本数 1
常见问题

LLM-Wiki-skills 是什么？

LLM-powered personal wiki knowledge base system. Use this when user wants to build and maintain a persistent wiki using LLMs following the LLM Wiki pattern (... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 186 次。

如何安装 LLM-Wiki-skills？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install llm-wiki-skills」即可一键安装，无需额外配置。

LLM-Wiki-skills 是免费的吗？

是的，LLM-Wiki-skills 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。

LLM-Wiki-skills 支持哪些平台？

LLM-Wiki-skills 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 LLM-Wiki-skills？

由 1knownothing（@1knownothing）开发并维护，当前版本 v1.0.0。

推荐 Skills
self-improving agent
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
⬇ 463815 · by pskoett
Skill Vetter
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
⬇ 259802 · by spclaudehome
Polymarket
Query Polymarket prediction markets. Check odds, find trending markets, search events, track price movements.
⬇ 232503 · by joelchance
Self-Improving + Proactive Agent
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
⬇ 200680 · by ivangdavila
Github
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
⬇ 191345 · by steipete
ontology
Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linkin...
⬇ 190333 · by oswalpalash

💬 留言讨论