Skip to content

About

Personal LLM-wiki workspace

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

Personal LLM Wiki

A general-purpose, LLM-maintained personal knowledge base, built on the pattern from Karpathy's LLM-wiki gist.

Instead of re-retrieving raw documents at query time, an LLM agent incrementally builds and maintains a persistent wiki. Knowledge compounds across sessions: you supply sources and good questions; the agent does the summarizing, cross-referencing, and consistency bookkeeping.

The wiki is domain-general — it can hold knowledge about anything (people, ideas, organizations, events, …). Deep, specialized subjects are added as domain forks (e.g. AI) that layer extra rules and page types on top of the general core, without polluting it.

Storage is plain markdown + git — readable in any editor, no special tooling required.

How it works — three layers

  1. Raw layer (raw/) — captured source material, in two stages:
    • raw/inbox/ — originals in native form (yyyy-mm-dd-name.{html,pdf,…}). Immutable, and git-ignored (local staging only).
    • raw/sources/ — a faithful markdown copyedit, 100% true to the original. The canonical, version-controlled record. The only edits allowed: Canadian-English spelling, meaning-preserving rephrasing, grammar fixes, and repairing broken lists/tables/formulas.
  2. Wiki layer (wiki/) — extracted, cross-referenced knowledge in typed folders. All interpretation and synthesis lives here.
  3. Schema layer — CLAUDE.md defines the conventions and workflows. AGENTS.md points non-Claude agents to it.

Structure

CLAUDE.md          schema + workflows (the agent's instructions)
AGENTS.md          cross-agent pointer to CLAUDE.md
index.md           catalog of all wiki pages, grouped by type
log.md             append-only activity log (ingests, queries, lints)

raw/
  inbox/   (git-ignored)   originals + assets
  sources/                 faithful markdown copyedits + renamed assets

wiki/                      general-purpose core types
  people/  orgs/  places/  events/  concepts/
  entities/  topics/  queries/  synthesis/

domains/
  ai/        example domain fork (rules + material only; pages go to root wiki/)

Workflows

The agent operates through three commands, defined in CLAUDE.md:

  • Ingest — add <url-or-file>: capture the original → produce a faithful copyedit → extract knowledge into wiki/, cross-linking pages and updating index.md and log.md.
  • Query — ask <question>: search the wiki and synthesize a cited answer, optionally filing it back into wiki/queries/.
  • Lint — lint: scan for contradictions, stale claims, orphaned pages, broken cross-references, and knowledge gaps.

Domain forks

A domain fork (domains/<name>/) bolts on a deep sub-domain without polluting the general schema. It holds rules and reference material only — a SCHEMA.md the agent layers on top of CLAUDE.md during ingest. It has no wiki pages of its own; every page it produces goes to the root wiki/, with its extra types becoming additional wiki/ subfolders. See domains/ai/SCHEMA.md for a worked example.

Conventions

  • Slugs: lowercase-kebab-case filenames.
  • Dates: ISO yyyy-mm-dd.
  • Spelling: Canadian English.
  • Faithfulness: raw/sources/ stays 100% true to the original; synthesis lives only in wiki/.
  • Cross-references: bidirectional relative markdown links.

Getting started

  1. Point your agent at this repo (it reads CLAUDE.md automatically).
  2. add <url> to ingest your first source.
  3. ask <question> to query what you've built.

The full design rationale lives in docs/superpowers/specs/.

Licence

MIT.

About

Personal LLM-wiki workspace

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors