Draft syncs shared AI agent context across teams via git
Draft is a free, open-source tool that automatically syncs AI agent context across a team by pulling and pushing edits to a shared private git repo, with human-in-the-loop verification.
Score breakdown
Draft introduces a git-backed, human-verified context layer that lets multiple agents and team members share the same AI session context, replacing ad-hoc per-user context management with a collaborative, auditable workflow.
- 01Draft is free and open source, hosted at github.com/idodekerobo/draft.
- 02It syncs AI agent context to a shared private git repo automatically.
- 03Claude Code analyzes agent sessions, Slack threads, and Granola notes/transcripts to build context.
Draft is a free, open-source tool created by idodekerobo to address the challenge of using AI agents collaboratively within a team. The core mechanism involves Claude Code analyzing agent sessions, Slack threads, and Granola notes and transcripts, then automatically syncing context edits to and from a shared private git repository. Because the process is human-in-the-loop, team members can review and verify what context is being added before it propagates to the rest of the team.
Draft ships as a macOS app, a CLI, and agent skills so that agents can work with it natively.
The shared context can be injected into multiple agents, with current support for Claude Code, Codex, OpenClaw, and Hermes. Draft ships as a macOS app, a CLI, and agent skills so that agents can work with it natively. The project is open source and hosted on GitHub.
Key facts
- 01Draft is free and open source, hosted at github.com/idodekerobo/draft.
- 02It syncs AI agent context to a shared private git repo automatically.
- 03Claude Code analyzes agent sessions, Slack threads, and Granola notes/transcripts to build context.
- 04The workflow is human-in-the-loop, letting users verify what is added to context.
- 05Shared context can be injected into Claude Code, Codex, OpenClaw, and Hermes.
- 06Draft ships as a macOS app, a CLI, and agent skills for native agent integration.
Topics
Summary and scoring are generated automatically from the original article. We always link back to the publisher and never republish images or paywalled content. Last processed Jun 18, 2026 · 10:40 UTC. How this works →