Back to directory

LLM Wiki

A Claude Code plugin that turns research, ideas and decisions captured while working into an interlinked personal wiki with search, automatic research on a missed lookup, and a browsable web interface.

LLM Wiki

LLM Wiki automatically saves research, ideas, decisions and findings to a wiki as the user works, checking the wiki first and researching and ingesting a topic automatically when it is not already covered, using whichever tools such as web search or an MCP connection are available. It supports block references and page embeds, a backlinks panel with unlinked-mention detection, and a query language for filtering pages by frontmatter fields, plus a nine-tier freshness system that tracks how quickly different kinds of content go stale.

Its bundled web interface offers a Wikipedia-style browsable site in four themes, an interactive knowledge graph, a canvas view for arranging pages spatially, a split-pane markdown editor with AI assistance, a spaced-repetition review mode and a chat sidebar that answers questions using retrieval over the wiki. A maintenance pass lints broken links, merges duplicate pages, upgrades confidence levels and flags stale content, and changes are committed to git automatically.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.