Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/frankchu91/mindbase-llm-wikiWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/plugins/frankchu91/mindbase-llm-wiki/plugin)<a href="https://agentmods.dev/plugins/frankchu91/mindbase-llm-wiki/plugin"><img src="https://agentmods.dev/badge/plugins/frankchu91/mindbase-llm-wiki/plugin/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/plugins/frankchu91/mindbase-llm-wiki/plugin"><img src="https://agentmods.dev/badge/plugins/frankchu91/mindbase-llm-wiki/plugin.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
mb scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
{
"name": "mb",
"description": "Karpathy's LLM-Wiki pattern as a Claude Code plugin. Build and maintain a persistent, AI-tended research wiki from sources, capture daily thoughts, query what you know, lint for orphans and contradictions. Operations: /mb:init, /mb:contribute, /mb:build, /mb:ask, /mb:lint, /mb:status, /mb:research, /mb:daily-brief, /mb:migrate.",
"version": "0.4.3",
"author": {
"name": "Haobing Chu",
"email": "[email protected]"
},
"homepage": "https://github.com/frankchu91/mindbase#readme",
"repository": "https://github.com/frankchu91/mindbase",
"license": "MIT",
"keywords": [
"wiki",
"second-brain",
"karpathy",
"research",
"knowledge-base",
"obsidian-compatible",
"markdown"
]
}
What it installs
The manifest is a name and a version. 1 skill, 12 commands, 5 agents, 1 hook, 1 MCP server travel with it, and installing the plugin installs all of them — 721 tokens a session between them. Each is measured on its own page, and each can be installed alone.
- Skill mindbase A 105 tokens
- Command contribute A 53 tokens
- Command build A 26 tokens
- Command init A 28 tokens
- Command research A 32 tokens
- Command lint A 37 tokens
- Command new-project A 20 tokens
- Command status A 32 tokens
- Command ask A 36 tokens
- Command daily-brief A 33 tokens
- Command load A 25 tokens
- Command migrate A 38 tokens
- Command export A 23 tokens
- Agent curator A 49 tokens
- Agent migrator A 48 tokens
- Agent builder A 39 tokens
- Agent contributor A 43 tokens
- Agent researcher A 54 tokens
- Hook SessionStart A not measured
- MCP server mb A not measured
What ships with it
1 file beside plugin.json in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 22 lines scan A 4147889efe46
mb is a plugin published in the GitHub repository frankchu91/mindbase-llm-wiki (130 stars, last pushed 6d ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other plugins, from other repositories
llm-wiki
Self-maintaining LLM wiki (Karpathy pattern): recall, observe, retro, capture sources, ensure pages, lint, and ingest. 15 tools over MCP.
llm-wiki
Persistent L1/L2 knowledge base in Logseq or Obsidian: ingest, query, prune, lint, status.
llm-wiki
Build a shared Markdown knowledge base with local hybrid search and an optional evidence-linked workflow to capture task lessons, evaluate skill improvements, and apply or roll back tested changes.
llm-wiki-manager
Personal LLM-managed wiki: ingest sources, cross-reference pages, query with citations, and lint your markdown knowledge base.
gnosis
Zero-config knowledge base MCP server — search, manage, and embed documentation (SQLite default, PostgreSQL optional).
llm-wiki-pm
PM knowledge base skill implementing Karpathy's LLM Wiki pattern. Ingest sources, query a compounding markdown wiki, supersede with auto-link rewrite, crystallize transcripts into decision digests. Bundles wiki search (obsidian-mcp).