You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
CLI hangs at startup/resume after "Managed MCP policy application exceeded its settings settlement deadline" #5341
Copilot CLI 1.0.105 on macOS (Apple Silicon), signed in with an Enterprise Managed User account whose organization policy sets forceRemoteSettingsRefresh.
The CLI sometimes hangs indefinitely at startup or on --resume. It has happened repeatedly over the past few days.
The server policy is fetched successfully (serverFetchFailed=false).
About 30 s later the log shows Managed MCP policy application exceeded its settings settlement deadline; independent enforcement continues (logged twice).
Nothing is logged after that. Only 1 or 2 of the 11 configured MCP servers (5 stdio, 6 HTTP) are started, the UI never becomes usable, and after about 2 minutes all connections to github.com are closed. I waited more than 3 minutes before killing it.
On resume, the session also logs: Plugins required by your organization admin could not be checked… Error: Managed settings could not be confirmed.
It happens intermittently with the same configuration. Starting with every --disable-mcp-server flag usually works, but it isn't tied to a particular server, because turning them back on one by one didn't reproduce it.
CLI processes left running from closed terminals seem to make it more likely, but stopping them doesn't always fix it.
More data: this reliably reproduces when resuming a session from inside a running CLI (/resume). In one process, a fresh session worked fine for ~25 min; then on resume the log shows Failed to reload managed hooks after enabling managed settings {"error":"no hook session for session id <resumed-id>"}, followed ~30 s later by Managed MCP policy application exceeded its settings settlement deadline, then no further progress (waited 6+ min). Same sequence seen on an earlier --resume <id> hang.
Activity
npeltier commentedon Nov 4, 2026
More data: this reliably reproduces when resuming a session from inside a running CLI (
/resume). In one process, a fresh session worked fine for ~25 min; then on resume the log showsFailed to reload managed hooks after enabling managed settings {"error":"no hook session for session id <resumed-id>"}, followed ~30 s later byManaged MCP policy application exceeded its settings settlement deadline, then no further progress (waited 6+ min). Same sequence seen on an earlier--resume <id>hang.