Search everything...

Stats

Actions

Available In

Help us improve

Share bugs, ideas, or general feedback.

llm-wiki - Claude Code Plugin | ClaudePluginHub

Plugin

llm-wiki

Name: llm-wiki
Author: marco3939

By marco3939

Turn Claude Code + Obsidian into a second brain. The LLM incrementally ingests sources into a persistent, interlinked markdown wiki — building entity/concept/source pages, flagging contradictions, maintaining an index and log. Knowledge compounds instead of being re-derived by RAG on every query. Inspired by Karpathy's LLM Wiki gist. Ships SKILL, 3 sub-agents, 5 slash commands, 8 Python tools (stdlib only), full vault templates, and cross-tool compatibility (Claude Code, Codex CLI, Cursor, Antigravity, OpenCode, Gemini CLI).

npx claudepluginhub marco3939/claude_skill --plugin llm-wiki

Popularity

Stars

Above avg

Med: 0·Avg: 271

What's Inside

Slash Commands5

wiki-ingest

/wiki-ingest

Ingest a source file from raw/ into the LLM Wiki — read, discuss, write summary page, update cross-references across 5-15 pages, regenerate index, append to log. Usage /wiki-ingest <path-to-source>

wiki-init

/wiki-init

Bootstrap a fresh LLM Wiki vault with the three-layer structure, schema files, and starter templates. Usage /wiki-init <path> --topic "<topic>" [--tool all|claude-code|codex|cursor|antigravity]

wiki-lint

/wiki-lint

Run a health check on the LLM Wiki vault — mechanical checks (orphans, broken links, stale pages, missing frontmatter, log gap, duplicates) plus semantic checks (contradictions, cross-reference gaps, concepts missing their own page). Outputs a markdown report with suggested actions. Usage /wiki-lint [--stale-days N] [--log-gap-days N]

wiki-log

/wiki-log

Show recent entries from the LLM Wiki log (wiki/log.md). Uses the standardized

wiki-query

/wiki-query

Query the LLM Wiki — reads index.md first, drills into 3-10 relevant pages, synthesizes an answer with inline [[wikilink]] citations, and offers to file the answer back as a new comparison or synthesis page. Usage /wiki-query "<question>"

Agents3

wiki-ingestor

/wiki-ingestor

Dispatched sub-agent that ingests a new source into an LLM Wiki vault. Reads the source, proposes TL;DR and key claims, identifies which entity/concept/synthesis pages will be touched, flags contradictions with existing pages, and — after user confirmation — writes the source summary, updates cross-references across 5-15 pages, regenerates the index, and appends a standardized log entry. Spawn when the user says "ingest this", "add this paper/article/book to the wiki", or drops a file into raw/.

wiki-librarian

/wiki-librarian

Dispatched sub-agent that answers queries against an LLM Wiki vault. Reads index.md first, drills into 3-10 relevant pages across categories, synthesizes an answer with inline [[wikilink]] citations, and offers to file the answer back into the wiki as a new comparison or synthesis page. Spawn when the user asks a substantive question the wiki might answer, says "what does the wiki say about X", "compare A and B across my sources", or wants to explore a topic.

wiki-linter

/wiki-linter

Dispatched sub-agent that runs a periodic health check on an LLM Wiki vault. Runs mechanical checks via scripts (orphans, broken links, stale pages, missing frontmatter, duplicate titles, log gaps), does semantic checks (contradictions, stale claims, cross-reference gaps, concepts missing their own page), and produces a markdown report with suggested actions. Spawn weekly, after batch ingests, or when the user says "check the wiki" / "lint my wiki" / "audit the vault".

Skills1

llm-wiki

/llm-wiki

Use when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.

README

marco3939/claude_skill — Portable Claude Code skills template

A personal, cloud-shareable Claude Code skills repository. Combines my own skills with a security-audited mirror of the alirezarezvani/claude-skills marketplace (313+ skills across 59 plugins).

What's in here

My skills (root level): algorithmic-art/, claude-api/, mcp-builder/, skill-creator/, pdf/, docx/, pptx/, xlsx/, webapp-testing/, theme-factory/, frontend-design/, canvas-design/, claude-canva-design/, slack-gif-creator/, web-artifacts-builder/, brand-guidelines/, doc-coauthoring/, internal-comms/, artfact-quiz/ — 19 hand-curated skills.
Community marketplace (categorized folders): engineering/, engineering-team/, c-level-advisor/, marketing-skill/, compliance-os/, ra-qm-team/, product-team/, business-growth/, finance/, productivity/, project-management/, research/ — 59 installable plugins via the bundled marketplace.

Setup on a new machine

Prerequisite: install Claude Code CLI

# needs Node.js 18+
npm install -g @anthropic-ai/claude-code
# or download from https://claude.com/claude-code

macOS / Linux

git clone https://github.com/marco3939/claude_skill.git
cd claude_skill
bash scripts/setup-user-memory.sh   # one-time: adds discovery protocol to ~/.claude/CLAUDE.md + SessionStart hook to ~/.claude/settings.json
claude

Windows (PowerShell)

git clone https://github.com/marco3939/claude_skill.git
cd claude_skill
.\scripts\setup-user-memory.ps1     # native PowerShell version of the setup
claude

The .ps1 script is compatible with Windows PowerShell 5.1 and PowerShell 7+. If you get an execution-policy error, run Set-ExecutionPolicy -Scope CurrentUser RemoteSigned once.

Then, inside Claude Code (once per machine)

/plugin marketplace add marco3939/claude_skill

Install plugins per use case:

/plugin install engineering-skills@claude-code-skills
/plugin install c-level-skills@claude-code-skills
# ...etc — install whichever plugins you need

After running the setup script, every future Claude Code session — even in unrelated repos — starts with a compact marketplace summary injected and the discovery protocol active. Claude will proactively suggest the right plugin for whatever you're working on; you confirm, install, proceed.

How AI uses this repo

The CLAUDE.md at root tells Claude Code to check the marketplace first before writing code. When you ask it to do something domain-specific (build a React app, audit GDPR compliance, write a CFO memo), Claude will scan .claude-plugin/marketplace.json and suggest the relevant skill — you confirm, install, and proceed.

This is lazy-loading: you don't pre-install 59 plugins; Claude proposes what's relevant when relevant.

Security audit

The vendored upstream was audited on 2026-05-19 in an isolated sandbox before merging. Findings: no backdoors, no malicious payloads, no data exfiltration. Full checklist:

✅ No curl ... | bash patterns in install scripts
✅ No base64 -d | sh or eval/exec of dynamic strings
✅ Hooks are defensive (local-only warnings & logs)
✅ No suspicious binaries, no obfuscated code
✅ Network access limited to user-provided URLs in SEO/sitemap utilities
✅ MIT-licensed with clear ownership (Alireza Rezvani)
✅ Project ships its own AUDIT_REPORT.md, VirusTotal CI scan, and skill-security-audit workflow

What was not vendored: upstream's .git/, .github/workflows/ (kept this repo's own CI), .mcp.json (the tessl MCP server is a tooling-time dependency, not needed for end users), and the *.zip archives (they duplicate the unpacked skill folders).

Attribution

Vendored skills: MIT © 2025 Alireza Rezvani — see LICENSE-CLAUDE-SKILLS and NOTICE.md.
My own skills: see LICENSE.txt inside each skill folder.

Updating from upstream

To pull newer upstream content later:

git clone --depth 1 https://github.com/alirezarezvani/claude-skills.git /tmp/upstream
# review the diff, then selectively rsync/copy folders you want

Don't blind-copy without re-auditing — upstream may add new skills or hooks that need fresh review.

Similar Plugins

c4-architecture

35.6k·179·

Comprehensive C4 architecture documentation workflow with bottom-up code analysis, component synthesis, container mapping, and context diagram generation

2mo

v1.0.0

wshobson

ecc

211.1k·3.7K·

Harness-native ECC plugin for engineering teams - 64 agents, 261 skills, 84 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses

today

v2.0.0-rc.1

affaan-m

context7-plugin

55.5k·339·

Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.

3mo

vctx7@0.3.9

upstash

octo

3.5k·106·

v9.42.3 — Patch release for cursor-agent smoke checks in untrusted workspaces. Run /octo:setup.

v9.42.3

v2.2.2

marco3939

Help us improve

Find plugins for your project

Help us improve

llm-wiki

Popularity

Health & Quality

Confidence

What's Inside

README

marco3939/claude_skill — Portable Claude Code skills template

What's in here

Setup on a new machine

Prerequisite: install Claude Code CLI

macOS / Linux

Windows (PowerShell)

Then, inside Claude Code (once per machine)

How AI uses this repo

Security audit

Attribution

Updating from upstream

Similar Plugins

c4-architecture

ecc

context7-plugin

octo

More by marco3939

terraform-patterns

research-summarizer

helm-chart-builder

code-tour

agenthub

Help us improve

unity-dev-toolkit

claude-md-management