Operator Mode gives Claude Code a lightweight execution memory layer.
It creates a single file — PROJECT_STATUS.md — inside a repo and teaches Claude to use it as persistent working memory so work resumes cleanly after context resets, without adding project-management overhead.
Claude is powerful, but long-running work can lose continuity as context fills.
Operator Mode solves that by:
- Maintaining a single high-signal status file
- Preserving intent, progress, and next actions across sessions
- Avoiding heavy frameworks, tickets, or scaffolding
It's not project management. It's continuity.
Clone this repository and run the installer:
git clone https://github.com/jeffwhitlock/operator-mode.git
cd operator-mode
chmod +x install.sh
./install.shRestart Claude Code if it is already running.
Inside any repository, run:
/operator-new
This initializes Operator Mode for that repo.
A single file at the repo root:
PROJECT_STATUS.md
Claude maintains this file as working memory throughout the session.
cd operator-mode
./uninstall.shThis removes the installed skill from ~/.claude/skills/.
(Project files are never modified or removed.)
Operator Mode is designed to be:
- Minimal
- Human-readable
- Repo-local
- Easy to remove
- Invisible once initialized
No databases. No services. No lock-in. Just a smarter way for Claude to stay oriented while you work.
MIT