# Totem

A file-based toolkit that keeps a project's coding lessons and rules inside the repository, compiles them into a deterministic, offline lint rule set, and exposes a queryable index of past decisions for any AI coding agent.

Totem addresses AI coding agents repeating the same architectural mistakes because each session starts from zero. It stores lessons, rules, and context as plain Markdown files in the repository, underneath whichever coding agent is used, so what one session learned is available to the next.

`totem lint` compiles the stored rules into a deterministic, offline linter with zero LLM calls and a cost that scales with the size of a diff rather than project history. A local, rebuildable knowledge index is derived from the files for querying, and separate compiler and review commands are LLM-powered and opt-in.

It integrates with Claude Code, Gemini CLI, GitHub Copilot, JetBrains Junie, and other editors, with more integrations in progress.

Website: https://github.com/mmnto-ai/totem