Agent Memory Kit


Memory Kit

**One memory for every coding agent.** **Plain markdown in your repository — injected on Claude Code, Cursor and OpenCode, followed by Codex and Copilot, readable by anything. Dated, capped, audited: it doesn't rot.**

[](https://github.com/awrshift/agent-memory-kit/releases) [](LICENSE) [](docs/specs/README.md)

*"I wake up already knowing where we left off."* — the agent this kit builds.

Install

In Claude Code, into a repository you already have:

/plugin marketplace add awrshift/agent-memory-kit
/plugin install memory-kit@memory-kit
/memory-kit:setup

Work as usual; say `/memory-kit:close-session` when you're done. That's the whole loop — no database, no service, no extra cost. Setup reads what your repo already has and **proposes before writing anything**; your `CLAUDE.md` stays yours. Codex and Cursor install from the same repository — see [Works with your agents](#works-with-your-agents).

Upgrading later — two commands, and the second needs the full id

claude plugin marketplace update memory-kit     # refresh the catalog
claude plugin update memory-kit@memory-kit      # the bare name resolves to nothing

Restart the session to apply. Upgrades never touch your repo — memory, handoffs and knowledge live in your files, not in the plugin.

The problem

Every session starts from zero: yesterday you locked the brand voice, today you explain it again. What your tools do remember is scattered — Claude Code's notes in one silo, Cursor's in another, none of it in your repo, none of it yours. And memory