Problem
relayfile integration connect <provider> establishes the connection (prints <provider> connected) but then blocks indefinitely waiting for the initial sync / keeping the mount active (relayfile-cli/main.go:1713 "Waiting for … initial sync", :1737 "Waiting for … sync status…", :1654 "keep this command running while the mount is active"). For a user whose goal is just to connect an integration (e.g. so agentworkforce deploy can resolve the persona's declared integrations), this looks like a hang and conflates three things: connect, mount, and initial-sync-wait.
Context
This blocking-on-sync was added deliberately (trajectory traj_v0l48chjqje6) in response to earlier feedback that the CLI "looked stuck" after connecting. But it overcorrected: now connect doesn't return at all until the (potentially long) initial sync completes, and it also makes agentworkforce deploy's integration-connect step block the same way.
Proposed change
integration connect returns once the connection is established + persisted (Nango connection exists).- Mounting + initial-sync-wait becomes opt-in: a
--mount/--watch flag on connect, or a separate relayfile mount <provider> command.
- Update the tests that currently assert the always-wait behavior (
main_test.go:921-925) to match the decoupled contract.
- Knock-on:
agentworkforce deploy integration-connect becomes non-blocking.
Priority
Low — workaround is Ctrl-C after " connected" (connection persists), or agentworkforce deploy --no-connect once integrations are connected. Filed from the overnight pr-reviewer E2E work where the slack connect appeared to stall.
Problem
relayfile integration connect <provider>establishes the connection (prints<provider> connected) but then blocks indefinitely waiting for the initial sync / keeping the mount active (relayfile-cli/main.go:1713"Waiting for … initial sync", :1737 "Waiting for … sync status…", :1654 "keep this command running while the mount is active"). For a user whose goal is just to connect an integration (e.g. soagentworkforce deploycan resolve the persona's declared integrations), this looks like a hang and conflates three things: connect, mount, and initial-sync-wait.Context
This blocking-on-sync was added deliberately (trajectory
traj_v0l48chjqje6) in response to earlier feedback that the CLI "looked stuck" after connecting. But it overcorrected: nowconnectdoesn't return at all until the (potentially long) initial sync completes, and it also makesagentworkforce deploy's integration-connect step block the same way.Proposed change
integration connectreturns once the connection is established + persisted (Nango connection exists).--mount/--watchflag onconnect, or a separaterelayfile mount <provider>command.main_test.go:921-925) to match the decoupled contract.agentworkforce deployintegration-connect becomes non-blocking.Priority
Low — workaround is Ctrl-C after " connected" (connection persists), or
agentworkforce deploy --no-connectonce integrations are connected. Filed from the overnight pr-reviewer E2E work where the slack connect appeared to stall.