HarnessCommunity★ 1k

My Claude Code Setup

Updated Jul 7, 2026

A starter template by centminmod that brings a Cline-inspired memory bank to Claude Code. Provides three updated CLAUDE.md template variants following Anthropic best practices (progressive disclosure, dual-memory architecture) plus slash commands for security audits, architecture review, usage cost analysis, and memory updates. Run /init once to populate the memory bank; use /update-memory-bank at session end to persist sprint state.

View on GitHub

What it does

/init
Analyze the codebase and populate the memory bank files at the start of a project.
/update-memory-bank
Write the current sprint state back to CLAUDE.md and memory bank files at session end.
/security-audit
Run an OWASP-guided security scan categorizing findings by Critical, High, Medium, and Low severity.
/secure-prompts
Detect prompt injection attacks and malicious instructions in file content or pasted text.
/create-release-note
Generate customer-facing and technical release notes from recent git commits, interactively or by commit count.
/ccusage-daily
Parse ccusage output into a structured daily cost, token usage, and cache efficiency report.

Files (1)

OVERVIEW.mdprimary · markdown · 2.1 KB

## What it does

A Claude Code starter config by centminmod that implements a Cline-inspired dual-memory architecture so Claude picks up exactly where the last session left off. Ships three CLAUDE.md template variants (progressive disclosure, dual-memory, standalone behavioral rules) updated to follow Anthropic best practices, a migration guide for existing users, and 20+ slash commands organized across security, architecture, documentation, and memory categories. Optional integrations for macOS Terminal-Notifier, Cloudflare/ClerkOS, and Convex database docs are included in supplementary files.

## Key commands

- /init — analyze the codebase and populate the memory bank from scratch
- /update-memory-bank — write current sprint state back to CLAUDE.md and memory files
- /security-audit — OWASP-guided security scan of the codebase with severity categories
- /secure-prompts — detect prompt injection attacks in file content or pasted text
- /check-best-practices — language-specific code quality analysis with before/after examples
- /create-release-note — generate customer-facing and technical release notes from git commits
- /ccusage-daily — parse ccusage output into a daily cost and token usage summary
- /apply-thinking-to — apply Anthropic extended thinking patterns to an existing prompt file
- /convert-to-todowrite-tasklist-prompt — convert a verbose workflow into parallel-subagent tasks (60-70% speed improvement)
- /cleanup-context — remove duplicate and obsolete memory bank content (15-25% token reduction)

## Install

Clone the repo and copy files into your project directory:

    git clone https://github.com/centminmod/my-claude-code-setup.git
    cp CLAUDE.md /path/to/your/project/
    cp -r .claude/ /path/to/your/project/

Install recommended CLI tools on macOS:

    brew install ripgrep fd jq

Then open Claude Code in your project and run /init to populate the memory bank.

## When to use

Use this when you want a production-ready CLAUDE.md setup with persistent memory across sessions, a suite of security and architecture slash commands, and a structured workflow for session start and end — without building it yourself.