Read Bridge Job Status

Require one job identifier. Run python "${CLAUDE_PLUGIN_ROOT}/bin/bridge_call.py" status --job-id "<job-id>"; always pass the same --workspace value the originating detached call used; a different or omitted value can resolve to another default workspace and report missing for a job still running elsewhere. Return JSON status unchanged — passthrough only; leave interpretation to /bridge:result. One call per check, never wrapped in a sleep, ScheduleWakeup, ListAgents, or Monitor loop; any state other than queued or running ends the wait.

Exception: cancel_requested does not end the wait — the supervisor and any write-capable child are still being torn down; the final state becomes cancelled. Terminal states: finished, cancelled, failed (the supervisor itself errored), stalled (supervisor gone, or still queued after 120 s), missing. For failed or stalled route the caller to /bridge:result; an implement job may already have edited files. When the originating call omitted --workspace, pass the absolute directory it ran from, never . — a later cd makes . another workspace. A status=error object (exit 2) is a refused lifecycle call, such as a malformed job id: report its error, never retry unchanged.