What is centminmod/my-claude-code-setup?

my-claude-code-setup by centminmod is an MIT-licensed starter configuration for Claude Code: CLAUDE.md templates built as a memory-bank system, plus hooks and slash commands. It aims to help Claude Code retain context across many sessions, and ships several README styles and supplementary docs for Cloudflare, Convex and more.

Can I use centminmod/my-claude-code-setup commercially?

Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.

Is centminmod/my-claude-code-setup still maintained?

Yes. The repository last received commits 14 days ago.

What programming language is centminmod/my-claude-code-setup written in?

Mainly Python, according to GitHub's language statistics.

my-claude-code-setup: CLAUDE.md Memory Bank Template

At a glance

What is it?: my-claude-code-setup by centminmod is an MIT-licensed starter configuration for Claude Code: CLAUDE.md templates built as a memory-bank system, plus hooks and slash commands. It aims to help Claude Code retain context across many sessions, and ships several README styles and supplementary docs for Cloudflare, Convex and more.
Who is it for?: Use my-claude-code-setup if you want a considered starting point for a project's CLAUDE.md rather than a blank file. Copy the files into your project, pick one of the three modern templates, remove the macOS settings.json if you are not on macOS, and run /init so Claude Code populates the memory bank.
Can I use it commercially?: Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
Is it still maintained?: Yes. The repository last received commits 14 days ago.
What is it written in?: Mainly Python, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on September 15, 2026, and from our analysis. They are not legal advice.

DEEP OPEN-SOURCE ANALYSIS

Hysen Labs Editorial · September 15, 2026 · centminmod/my-claude-code-setup

What it is and the problem it addresses

my-claude-code-setup is a shared starter template of configuration for Claude Code: a project's starter settings, Claude Code hooks and slash commands, and, at its centre, a CLAUDE.md set up as a memory-bank system of files so Claude Code retains context better across many chat sessions. It is not a program you run; it is a set of files you copy into a project and adapt.

The problem it targets is context loss. A single CLAUDE.md tends to drift or grow unwieldy, and Claude Code forgets project specifics between sessions. The memory-bank approach structures that context into files the model reads and updates, so the project's decisions and state persist. The audience is Claude Code users, from beginners to advanced, who want a thought-through starting configuration instead of assembling one from scratch, and who are willing to prune the parts they do not need.

The CLAUDE.md templates and memory bank

The core of the repository is its CLAUDE.md. A May 2026 update modernised it with three new variants, CLAUDE-template-1.md, CLAUDE-template-2.md and CLAUDE-template-3.md, following what the README describes as official Anthropic best practices, including a dual-memory architecture, progressive disclosure and standalone behavioural rules. Existing users are given CLAUDE-migrate-to-new-template.md, which is itself an AI prompt: you feed it to Claude to migrate your current CLAUDE.md to one of the new templates while preserving your project-specific content.

The memory-bank idea is that CLAUDE.md instructs Claude Code to maintain a set of files that retain context over many sessions. The README shows the intended workflow in practice: after completing a task and committing, you tell Claude to update memory bank, and it refreshes those files. This is the feature that distinguishes the repo from a generic config, and it is worth understanding before adopting, because the value depends on actually running that update step as part of your routine.

Setting it up

The README's setup is a numbered list, and the essential steps are straightforward. First, copy the files from the repo into your project directory. Second, modify the templates and CLAUDE.md to your liking; note that .claude/settings.json expects Terminal-Notifier for macOS, and if you are not on macOS you can simply remove that file. Third, after launching Claude Code in your project for the first time, run /init so Claude Code analyses your codebase and populates the memory-bank files as CLAUDE.md instructs.

From there the README offers optional, recommended extras: installing VS Code and the Claude Code extension, setting up Git, and, for macOS, installing faster command-line tools that CLAUDE.md instructs the model to prefer:

bash

brew install ripgrep fd jq

Running /init after the copy is the step that actually populates the memory bank, so do not skip it. There is also an optional dev-container path that bundles Claude Code, Codex, Gemini CLI and other tools in an isolated Debian 12 container, which is the safer place to run permissive YOLO modes if you use them.

Supplementary docs and multiple README styles

The repository is modular: alongside the base CLAUDE.md there are supplementary reference docs you keep only if relevant. CLAUDE-cloudflare.md and a mini version help the model understand Cloudflare and ClerkOS platforms, and CLAUDE-convex.md supports building Next.js and React apps with a Convex backend on Cloudflare Pages. You reference whichever files apply from CLAUDE.md and delete the sections you do not use. There is also a GEMINI.md and an AGENTS.md, reflecting cross-tool use.

A nice touch is that the README itself comes in several styles, generated by asking Claude Code to rewrite it for different readers: the original README.md, a progressive-disclosure README-v2.md, a task-based README-v3.md organised around I want to..., and a dense reference-manual README-v4.md. That is a small but genuine demonstration of the project's own philosophy, and it means you can pick the explanation format that suits how you learn.

Higher quotas with Z.AI and other integrations

The README notes that Claude Code needs a paid Claude plan (Pro, or Max at higher tiers) with the usual quotas and rate limits, and points to Anthropic's official docs as required first reading. It also documents an alternative for more headroom: configuring Claude Code with Z.AI to get higher token-usage quotas and access to Z.AI's GLM models within Claude Code. The README includes a Z.AI invite/discount link, which is an affiliate promotion rather than a technical necessity, so treat it as optional.

The wider set of pointers, dev containers, Git Worktrees for the Claude Desktop app (with a .worktreeinclude file), and a reference on running Claude models with OpenClaw, shows the repo is as much a curated set of Claude Code practices as a template. That is useful, but it also means some content is opinion and some is promotion; take the configuration files as the substance and the external links as suggestions to evaluate on their own.

Licence, maintenance and how to treat it

The project is MIT licensed, so you can freely copy and adapt the templates, which is exactly how it is meant to be used. There are no tagged releases; the repository tracks changes through a CHANGELOG.md and dated notices in the README, and the last push was on 2026-09-02, so it is kept current with Claude Code's evolving best practices.

The right way to treat it is as a starting point, not a drop-in. Because it encodes one author's workflow, complete with macOS-specific notifier settings, platform docs for Cloudflare and Convex, and affiliate links, the first job after copying it is subtraction: remove the settings.json if you are not on macOS, delete the platform docs you do not use, and keep the CLAUDE.md template and memory-bank instructions that are the real value. Read Anthropic's official Claude Code documentation alongside it, since the templates assume that context.

Editorial conclusion

Use my-claude-code-setup if you want a considered starting point for a project's CLAUDE.md rather than a blank file. Copy the files into your project, pick one of the three modern templates, remove the macOS settings.json if you are not on macOS, and run /init so Claude Code populates the memory bank. Read Anthropic's official Claude Code docs alongside it.

Frequently asked questions

How do I configure Claude Code with this setup?

Copy the repo files into your project, adapt CLAUDE.md and the templates, remove .claude/settings.json if you are not on macOS, then run /init in Claude Code so it analyses your codebase and populates the memory-bank files as CLAUDE.md instructs.

What is the CLAUDE.md memory bank system?

It is a structure where CLAUDE.md tells Claude Code to maintain a set of files that retain project context across many sessions. You refresh them by telling Claude to update memory bank after completing a task, so context is not lost between chats.

Which CLAUDE.md template should I use?

The repo ships three modern variants (CLAUDE-template-1/2/3.md) following Anthropic best practices with progressive disclosure and dual-memory architecture. Existing users can feed CLAUDE-migrate-to-new-template.md to Claude to migrate their current CLAUDE.md while keeping project-specific content.

Official sources

Community notes