I've been using Claude Code for a few months and kept running into the same issues:
1. No enforcement of TDD or quality checks - Claude writes untested code if you let it
2. Context window fills up mid-feature, losing all the understanding Claude built up
3. Starting fresh sessions means re-explaining the codebase all over again
4. Setting up a good configuration takes time and effort to keep up with everything
I built Claude CodePro to address these for my own freelance work. Sharing it in case others find it useful.
---
TWO WAYS TO WORK
1. /spec "your feature" - For larger work - Creates an implementation plan as markdown - You review and approve before implementation - TDD enforcement via pre-edit hook - Verification step at the end
2. Quick Mode - Just chat - No planning overhead - Quality hooks still run - Good for bug fixes and small changes
---
ENDLESS MODE
One of the main features is automatic context continuation. When you hit ~90% of the 200K context window, the system:
- Saves current state to a continuation file
- Captures observations via Claude Mem (persistent memory MCP)
- Wrapper automatically triggers a session restart
- New session reads the saved state from hooks and files and continues
I've had features span multiple session handoffs without manually managing anything.
---
QUALITY HOOKS
Post-edit hooks run automatically:
- Python: ruff, basedpyright
- TypeScript: eslint, tsc, prettier
- All languages: QLTY
Pre-edit hook warns if you're modifying code without a failing test first.
---
DEV CONTAINER
Everything runs in a Dev Container. Advantages:
- Consistent environment across all machines and operating systems
- Complete isolation from your host system
- Complex toolchain (LSP servers, semantic search, linters) just works
- No manual installation of dependencies
- Easy onboarding - works immediately after container starts
- Reproducible setup every time
---
INCLUDED TOOLS
- Vexor: Local semantic code search
- Claude Mem: Persistent memory across sessions
- Context7: Library documentation retrieval
- Firecrawl: Web scraping
- agent-browser: Headless Chromium for browser automation
- mcp-cli: CLI for custom MCP servers
---
LINKS
GitHub: https://github.com/maxritter/claude-codepro
Website: https://claude-code.pro