Context OS

What Context OS does

Context OS is a Git-backed context and workflow layer that keeps durable project context in plain Markdown: who you are, what you are working on, decisions already made, and the workflows you want an agent to follow. Supported coding agents read and update it through reviewed handoffs instead of relying on chat history, project instructions and copied prompts that drift apart.

Claude Code, Codex, OpenClaw, OpenCode, Cursor CLI, Devin CLI and Hermes Agent have first-class adapters, and Cursor IDE support is experimental. The adapters route to shared repository state, and a deterministic lifecycle kernel turns reviewed setup, checkpoint and close requests into hash-checked proposals and receipts, while host-native memory stays separate. It includes a source-neutral migration workflow, an integration catalog that documents install scope and side effects, session handoffs, staleness checks and decision logs. It does not scrape accounts or install integrations automatically.

Key features

• Plain-Markdown context repository backed by Git
• First-class adapters for Claude Code, Codex, OpenClaw, OpenCode, Cursor CLI, Devin CLI and Hermes Agent
• Deterministic lifecycle kernel producing hash-checked proposals and receipts
• Source-neutral migration workflow for chats, instructions and memory exports
• Integration catalog with install scope, data access and confirmation gates
• Session handoffs, staleness checks and decision logs

When to use it

• Hand work from Claude Code to Codex or Hermes with the same project context
• Move useful context out of chat history into reviewable files
• Keep decisions and workflows versioned in a private Git repository

Who it is for: Developers who use more than one coding agent and want durable, reviewable context that does not live in any single tool.

How it fits with Hermes Agent

Hermes Agent is listed with first-class adapter support alongside other coding agents; Context OS is a general tool rather than a Hermes-specific one.

FAQ