Answers
Can my teammate see what Claude Code saved in my auto memory?
No. The Claude Code documentation I read says auto memory is notes Claude writes itself from your corrections and preferences, stored per project in a folder on your own machine, at ~/.claude/projects/<project>/memory/, and that these files are not shared across machines or cloud environments. All worktrees of the same repository on your machine share that one folder; a teammate's session does not read it. What a teammate does see is what is committed to the repository: the project CLAUDE.md is described as team-shared instructions through source control, while the user CLAUDE.md and the local, gitignored CLAUDE.local.md are only yours. So when Claude learns something on your machine that the whole team should rely on, it stays on your machine. Arroway is where that kind of thing goes so that the team's sessions all read it: an assistant records it as a proposal, a person approves it, and every tool on the project reads it before it acts. Use Arroway when what one session learned should reach the next person's session. The install page is the first step.
Last updated October 15, 2026
What stays on your machine and what reaches the team
Claude Code's documentation describes two kinds of memory. The files you write, CLAUDE.md, come in scopes: managed policy for an organization, user for you across projects, project for the repository, and a local file for you in one project. Only the project file travels with the repository, and it travels because it is committed. Auto memory is the other kind: Claude writes it itself, as a MEMORY.md index plus one file per topic, in a folder under your home directory. The documentation states that it is machine-local and not shared across machines or cloud environments, and that the index is loaded at the start of every session, up to its first 200 lines or 25 KB.
Why a teammate's session never sees it
Nothing in that design moves the notes from your folder to anyone else's. A teammate who clones the same repository gets the committed instruction files, not your auto memory. Two people working on one codebase can therefore hold different notes about it, each learned in a different session and neither reviewed by anyone. A note that Claude saved because you corrected it once is also a note nobody approved as what the team does. That is a property of where the file lives, not a defect to work around: auto memory is built to remember your sessions on your machine.
When to use Arroway
Use Arroway when something one session learned should be true for the next person's session as well, in whichever tool they use. You or an assistant records it on the project, and it stays a proposal until a person approves it. From then on the sessions on that project open by reading what is approved, each rule shows who approved it and the condition that ends it, and a newer rule names the one it replaces. Arroway does not collect the notes in your auto memory folder: what reaches the team is what is recorded on the project. The first step is the install page.
What it looks like in practice
Two developers, Ana and Bruno, work on the same repository. In a session on Tuesday, Ana corrects Claude Code twice: the integration tests only pass after the staging database is reset with the repository's reset script. Claude saves that as a note in Ana's auto memory, and her next sessions run the script first. On Thursday Bruno runs the same tests, and his sessions do not know about the reset. The tests fail, he asks Claude to fix them, and it spends an hour on a problem Ana already solved. Nothing was shared, because nothing was meant to be. If Ana records the reset step on the project and a person approves it, Bruno's next session reads it before it runs the tests, and so does a Codex or Cursor session of either of them. The people and the project in this example are illustrative; nothing here was measured on a real team.
Questions people ask about this
- Can I make a teammate read my Claude Code auto memory?
- Claude Code's documentation describes the notes as machine-local and does not describe sharing them with teammates. If a teammate needs to rely on something, the dependable route is to record it on the project, where it is approved and read by every tool the team uses.
- Does the project CLAUDE.md count as shared memory?
- The documentation describes the project CLAUDE.md as team-shared instructions delivered through source control, so teammates get it when they pull. It carries what someone wrote and committed. It does not carry what a session learned on one machine, and it does not show who approved a line or when it should stop applying.
- How do I try Arroway?
- Install it in the tool you already use — the install page has one route for Claude, ChatGPT, Codex, Cursor or any MCP tool — and sign in when the tool asks. From then on, sessions open by reading the project and close by recording what was decided.
Where this is verifiable
The memory page of the Claude Code documentation (code.claude.com/docs/en/memory), read on 15 October 2026 — Claude Code is described only as that page describes it on that date, and its features can change in any release. Product documentation on this site (How it works, Install), the answers on Arroway vs CLAUDE.md and on alternatives to CLAUDE.md for a team, and the sanction, supersession and expiry-condition rules in the sanctioned product spec. Everything described here about Arroway is behaviour the tools apply today, not roadmap. The developers in the example are illustrative.
https://www.arroway.app/en/answers/can-your-team-see-claude-code-auto-memory