What problem does it solve? Setting up a persistent, agent-maintained knowledge base inside a repository requires consistent directory structure, schema conventions, and documentation wiring that is tedious to create by hand. This Skill automates that one-time bootstrap so future sessions can ingest sources, query, and lint against a stable wiki. ## Core Features & Use Cases - Guided three-question interview: Captures the wiki's domain, expected source types (web articles, PDFs, transcripts, notes), and solo/team voice before writing anything. - Scaffolded layout: Creates .red/wiki/ (raw, pages, index.md, log.md) and renders .red/agents/wiki.md from bundled schema templates with placeholder substitution. - Repository integration: Appends .red/wiki/ to .gitignore and registers a ### Wiki block under ## Agent skills in CLAUDE.md or AGENTS.md. - Use Case: A researcher wants an incrementally maintained wiki of papers and articles in their repo. They run the init once, answer three questions, approve the preview, and immediately start ingesting sources with /wiki ingest. ## Quick Start Ask the agent to initialize the LLM Wiki in this repository and answer its three setup questions about domain, source types, and voice.
