Skip to content

release: multi-person fleets - #4

Merged
anianroid merged 2 commits into
mainfrom
staging
Oct 7, 2026
Merged

anianroid merged 2 commits into
mainfrom
staging

Conversation

@anianroid

Copy link
Copy Markdown
Contributor

Promote staging to main.

Made with Cursor

anianroid and others added 2 commits October 7, 2026 12:29
People listed in fleet.yaml share the fleet's subscriptions. With a Tailscale
auth key the dashboard identifies each person, so they can connect their own
ChatGPT and have their new chats billed to it.

Co-authored-by: Cursor <[email protected]>
feat: multi-person fleets with optional personal ChatGPT
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 7, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-07T07:11:44.116875Z d0e2956 PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d0e295672c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

services:
openclaw:
user: root
cap_add: ["NET_ADMIN"]

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Mount the TUN device for kernel-mode Tailscale

Whenever TS_AUTHKEY is set, the entrypoint starts a normal kernel-mode tailscaled, but this override grants only NET_ADMIN and never passes /dev/net/tun into the container. Tailscale documents that kernel networking requires both the device and capabilities, so the daemon exits and the new Tailscale-only Control UI is unreachable; either map /dev/net/tun or explicitly use userspace networking. Tailscale Docker parameters

Useful? React with 👍 / 👎.

Comment on lines +53 to +54
exec setpriv --reuid=node --regid=node --init-groups --inh-caps=-all \
/usr/local/bin/fleet-entrypoint gateway

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Grant node permission before running Tailscale commands

In Tailscale mode the daemon starts as root without configuring an operator, then these lines re-enter the script as node; consequently the subsequent tailscale up and OpenClaw's tailscale serve cannot manage the daemon. Tailscale states that only root can manage tailscaled unless another user is explicitly made its operator, so the login fails and the gateway cannot establish Serve. Tailscale operator permissions

Useful? React with 👍 / 👎.

Comment on lines +284 to +287
# One shared auth profile per subscription, labelled so people can pick it in chat.
for sub in subs:
provider = {"claude": "anthropic", "chatgpt": "openai"}[sub]
ops.append({"path": f"auth.order.{provider}", "value": [f"{provider}:fleet"]})

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Preserve existing shared auth profiles during upgrade

On an existing fleet, the old login commands stored credentials under automatically chosen profile IDs rather than anthropic:fleet or openai:fleet, but every ./fleet up now unconditionally restricts each provider to only the new ID. OpenClaw explicitly excludes stored profiles omitted from auth.order, so upgrading immediately disables previously working shared credentials until every subscription is logged in again; migrate the old profiles or avoid installing this order until the named profile exists. OpenClaw auth-order semantics

Useful? React with 👍 / 👎.

Comment on lines +155 to +156
if not ids and len(people) == 1:
ids = [i for i in [os.environ.get(env_var, "").lstrip("+")] if i]

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Keep owner environment IDs in multi-person allowlists

Fresh setup still stores the owner's channel IDs only in OWNER_* variables and leaves the generated people.owner channel fields blank. As soon as any teammate is added, len(people) == 1 becomes false, so those owner IDs are discarded; after the teammate IDs are filled in, only teammates are allowlisted and the owner unexpectedly loses DM access. Include the environment-backed owner identity even when other people are present, or copy it into the owner entry during setup.

Useful? React with 👍 / 👎.

Comment on lines +53 to +54
exec setpriv --reuid=node --regid=node --init-groups --inh-caps=-all \
/usr/local/bin/fleet-entrypoint gateway

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Drop privileges before starting gateway helpers

With TS_AUTHKEY set, the root invocation has already started the warden scheduler and fleet-board server before it reaches this exec, and those background children survive while the script re-enters as node. The second invocation starts both helpers again, causing every scheduled warden job to run twice while a second board supervisor repeatedly fails to bind port 8090; perform the privilege drop before launching either helper.

Useful? React with 👍 / 👎.

openclaw:
user: root
cap_add: ["NET_ADMIN"]
ports: !reset []

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Keep the fleet-board port published in Tailscale mode

The !reset [] override clears the entire openclaw.ports sequence, not just the gateway's 18789 mapping, so enabling Tailscale also removes the existing localhost mapping for fleet board on 8090. This breaks the SSH tunnel printed by ./fleet ui, while Tailscale Serve proxies only the gateway; override the ports selectively so 8090 remains published. Docker Compose reset semantics

Useful? React with 👍 / 👎.

@anianroid
anianroid merged commit a6e2242 into main Oct 7, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant