Browse / Software development / Productivity & assistant / claude-code-starter-kit
claude-code-starter-kit
rmbell09-lang
Production-tested Claude Code workspace setup: persistent memory, AGENTS.md, SOUL.md, 5 skills — get consistent AI agents in 10 min
Open rmbell09-lang/claude-code-starter-kit → On the boundary of the four rules.
agentsai-agentai-workspaceanthropicboilerplateclaudeclaude-codedeveloper-toolsopenclawpersistent-memoryproductivity
Does not say
no agent config and no instruction file at the root.
In its own words
Production-tested setup for building AI agents with Claude Code
Stop configuring. Start shipping.
This kit packages everything we use in production to run persistent, self-improving AI agents with Claude Code (OpenClaw). Drop these files into your workspace and your agent wakes up with memory, personality, and purpose — every session.
Quick Start (10 minutes)
- Copy
templates/AGENTS.md→ your OpenClaw workspace root - Copy
templates/SOUL.md→ same directory - Edit
templates/USER.mdwith your name + preferences, copy it too - Copy
memory/MEMORY.md→ your workspace root - Copy any
skills/you want →~/.openclaw/workspace/skills/ - Start a Claude Code session — your agent now has memory + personality
Full setup in guides/QUICK_START.md.
Why This Works
Most people use Claude Code as a stateless chatbot. We run it as a persistent agent that:
- Remembers everything across sessions via daily markdown notes
- Has consistent personality (no more robotic corporate replies)
- Knows who it's helping (reads USER.md before answering)
- Uses the right model for the task (not always Opus)
- Gets better over time (logs learnings, applies them)
This is the exact setup running in production right now. Not a demo.
Pricing
One-time purchase. No subscription, no license keys, no expiration. You get everything and own it forever.
Quoting 92% of the README. Read the rest on GitHub ↗
Read from the repository's git tree on 2026-08-17.