MemClaw Workspace Structure: Good vs Bad Examples
Good workspace structure makes Claude Code more effective. This guide shows good vs bad examples and explains how to organize your workspace for maximum queryability.
MemClaw Workspace Structure: What to Put Where
A MemClaw workspace is only useful if it's structured well. Too much noise and Claude can't find the signal. Too little and it's missing critical context. This guide covers the optimal structure for a MemClaw workspace used with Claude Code. Set up your workspace at memclaw.me → memclaw.me
The Core Sections
A well-structured workspace has five sections: Each section serves a specific purpose. Here's what goes in each.
Tech Stack
Short, factual. What Claude needs to know about your environment: Keep it to versions and deployment targets. Don't explain why — that goes in Decision Log.
Architecture Rules
The rules Claude must follow when suggesting code: These are constraints, not explanations. Short, imperative statements.
Current Sprint
What matters right now: Update this at the start of each sprint. The "not this sprint" section is as important as the P0 list.
Known Issues
Active bugs and constraints Claude should know about: Remove resolved issues or mark them resolved with the fix.
Decision Log
Why things are the way they are: This is the "why" that prevents Claude from suggesting rejected approaches.
Session History
What happened in recent sessions: Keep the last 10-15 sessions. Archive older ones if the workspace gets too large.
What NOT to Put in Your Workspace
- Source code (use the actual files)
- Credentials or API keys (use environment variables)
- Detailed technical specs (link to docs instead)
- Personal notes unrelated to the project
- Resolved issues older than 2 sprints
Getting Started
- Create a workspace at memclaw.me
- Add the six sections above
- Fill in Tech Stack and Architecture Rules first
- Add your current sprint goals
- Log your top 3 known issues Create your workspace → memclaw.me