LLM Wiki

# LLM Wiki

AI Summary

Purpose:

• Explains the LLM Wiki pattern used by this repository.

Key points:

• LLM Wiki is a persistent Markdown knowledge base maintained by AI agents.
• Raw sources are ingested, summarized, linked, and promoted into durable wiki pages.
• Obsidian can be used as the human browsing interface.
• Git provides version history and review.

Relevant when:

• Explaining this repository design.
• Processing raw sources.
• Deciding whether information belongs in ai/sources/, ai/wiki/, ai/weekly/, or ai/worklog/.

Do not read full document unless:

• You need details about the LLM Wiki operating model.

Linked documents:

• ai/wiki/concepts/token-efficient-reading.md
• ai/sources/index.md
• ai/weekly/index.md

Short Definition

An LLM Wiki is a structured Markdown knowledge base that AI agents maintain over time by reading raw sources, extracting key information, updating wiki pages, adding links, and flagging contradictions.

Key Points

• Raw documents are inputs, not the final knowledge base.
• The wiki is the compiled, durable memory layer.
• Agents should update existing pages instead of creating duplicate notes.
• Summaries and indexes reduce token usage.
• Weekly notes capture time-based work before durable knowledge is promoted.