Remote SSH reconnect loses persistent session after Mac sleep

[MrPhobos]

What happened?

After a brief network interruption caused by the local Mac sleeping, a Supacode Remote SSH (Beta) terminal does not reliably reattach to its persistent remote session.

I have observed two failure modes:

1. With Keeps SSH sessions alive across disconnects enabled, reconnection can loop with No user exists for uid 501 and SSH exit 255.
2. After disabling that option as a workaround, SSH reconnects successfully, but Supacode reports Reconnected without a persistent session; starting a fresh shell. The terminal scrollback remains visible, but the running Codex process is gone.

I expected Supacode to reconnect and reattach to the existing remote zmx session so the running process survives the short disconnect.

The network and SSH endpoint recover normally: in the second mode, SSH reconnects after one second. Remote zmx logs show that affected sessions received an explicit kill rather than merely losing their client, which suggests the session cleanup/reconnect lifecycle is terminating a session that is still needed.

Steps to reproduce

1. On a MacBook, open a repository on another Mac using Remote SSH (Beta), with Keeps SSH sessions alive across disconnects enabled.
2. Start a long-running interactive process such as Codex in the remote terminal.
3. Let the MacBook sleep briefly so the SSH/Tailscale connection drops, then wake it.
4. Observe repeated SSH exit 255 reconnect attempts with No user exists for uid 501.
5. Disable Keeps SSH sessions alive across disconnects, reconnect, and repeat the short sleep/disconnect.
6. SSH now reconnects, but Supacode falls back to a fresh shell instead of restoring the running process.

The exact sleep timing is intermittent, but this has happened repeatedly across multiple tabs in the same remote workspace.

Supacode version and build

0.10.8 (1785775286)

Latest published release checked on 2026-09-01: v0.10.8.

macOS version

• Local: macOS 26.5.1, Apple silicon
• Remote: macOS 13.7.8, x86_64

System locale

en_US@rg=nlzzzz

Mac hardware

MacBook Air, Apple M3, 16 GB

Relevant logs

Supacode terminal output:

Shared connection to remote host closed.
Connection failed (ssh exit 255). Retrying in 1s. Press Ctrl-C to stop.
Reconnected without a persistent session; starting a fresh shell.

Earlier persistent reconnect loop:

No user exists for uid 501
Connection failed (ssh exit 255). Retrying in 15s. Press Ctrl-C to stop.

Relevant remote zmx log lines from an affected session:

deleting socket file
kill received
shutting down daemon
sending SIGHUP to session process

Additional diagnostic detail:

Local zmx bundled with Supacode: 0.6.0
Remote /usr/local/bin/zmx: 0.7.0

The version mismatch may be relevant, but the clearest symptom is that the remote session receives a kill during the reconnect/cleanup sequence.

Before submitting

• I searched existing issues and this is not a duplicate.
• I am on the latest available version of Supacode.
• I understand pull requests are closed until the issue is marked ready.
• I agree to follow this project's Code of Conduct.

I am not currently planning to open a pull request.

[github-actions]

Thank you for filing this. A maintainer will review it and follow up here.

Please do not start any work or open a pull request until this issue is labeled ready. That label is a maintainer's go-ahead: a pull request opened against an issue that is not yet ready is labeled invalid until then. You will get a comment here the moment it is marked ready.

See CONTRIBUTING.md for the full flow.