contextkeeper 0.2.0

Zero model drift between AI agents. Universal session continuity protocol and CLI for Claude, GPT, Gemini, and any LLM.

contextkeeper

Zero model drift between AI agents. Universal session continuity protocol and CLI for Claude, GPT, Gemini, and any LLM.

Install

pip install contextkeeper

Commands

contextkeeper init

Initialize contextkeeper state files in the current project. Auto-detects project type, generates STATE_VECTOR.json and HANDOFF.md, and creates a .workbench config file.

cd my-project
contextkeeper init

Options:

• -p, --project — Project slug (default: directory name)
• -t, --type — Project type override
• --bridge — Bridge repo (e.g. yourname/workbench)

contextkeeper sync

Sync state files to your bridge repo.

contextkeeper sync

Options:

• --bridge — Bridge repo override
• --dry-run — Preview without pushing

contextkeeper status

Show status of all projects in the bridge repo.

contextkeeper status --bridge yourname/workbench

Options:

• --bridge — Bridge repo
• --json — Output as JSON

contextkeeper bootstrap

Generate a paste-ready bootstrap prompt for any AI chat.

contextkeeper bootstrap -p my-project --clipboard

Options:

• -p, --project — Project slug (required)
• --bridge — Bridge repo override
• --clipboard — Copy to clipboard

contextkeeper doctor

Check environment and configuration health.

contextkeeper doctor

How It Works

1. contextkeeper init creates structured state files in your project
2. contextkeeper sync pushes them to a central bridge repo on GitHub
3. contextkeeper bootstrap generates a prompt you paste into any AI chat
4. The AI reads your state files and has full context in under 60 seconds

Requirements

• Python 3.10+
• git

License

MIT