scidex-bridge.service was a remote file-management HTTP daemon (port
8899) that hadn't worked in years (3800+ failed restart cycles
logged on 2026-04-28; required SCIDEX_TOKEN env var that's not in
.env). The Orchestra fleet drives every change through git
worktrees + PRs now; nothing internal calls the bridge.
The systemd unit was stopped + disabled + removed on 2026-04-28.
SciDEX PR #1014 was opened the same day to remove the source files
+ every code reference. PR #1014 was closed without merging.
This task delivers the file-deletion + reference-cleanup half of
#1014 as a fresh PR. The runtime systemd cleanup already happened.
/home/ubuntu/scidex/bridge.py/home/ubuntu/scidex/scripts/bridge.pycli.py:SERVICES dict entry "bridge": {...} (line ~50)BRIDGE_SERVICE = "scidex-bridge" constant (line ~56)tar arg listing bridge.py (line ~870)sudo cp scidex-bridge.service in the bundle script (line ~857)
deploy/bootstrap/install.sh:scidex-bridge.service line in the services array (~line 139)
scidex/core/scidex_doctor.py:check_bridge() probe (line ~316-331). Replace with aAGENTS.md references (5 places):bridge.py (line ~65)bridge.py (line ~1056)Orchestra/orchestra/agent.py heartbeat / lease# bridge comments — those are unrelatedMessageBridge, event_bus, etc.).grep -rn "scidex-bridge\|bridge\.py" /home/ubuntu/scidex/cli.py<details>-tagged historicalls /home/ubuntu/scidex/bridge.py /home/ubuntu/scidex/scripts/bridge.pyNo such file.
scidex doctor (or scidex.core.scidex_doctor) runs withoutBridge probe.
gh pr for this work merges cleanly to main; no follow-upCleanup hygiene. The runtime damage is already neutralized (unit
removed). The repo state is just embarrassing: future operators
reading AGENTS.md still see "bridge.py" listed as a restart-sensitive
service.
docs/planning/2026-04-27-orchestra-session-summary.md.