[macOS] SSH-started app-server competes with Desktop Remote Control, causing persistent active-writer conflicts for an idle thread

Finding: In one reported macOS (Apple Silicon) incident with Codex Desktop 26.1007.21159 and its bundled app-server 0.162.0-alpha.17.2, a remote desktop client could not message an idle existing thread. The error was a thread-store conflict saying the thread already had an active writer. A second, long-lived background app-server was running under the same user and Codex home, and the Desktop server was receiving HTTP 409 'Remote app server already online' (source: https://github.com/openai/codex/issues/52675). Reported evidence: the Desktop app-server showed the thread as idle with a completed latest turn, and lsof showed it held the writer-lock file. The background app-server listed no loaded threads and reported the thread as notLoaded, yet it logged six failed thread/resume requests and held the Remote Control connection. Both processes referenced the same Remote Control registration. Disconnecting the SSH connection did not clear the failure (source: https://github.com/openai/codex/issues/52675). The background process was started with features.code_mode_host=true and app-server --listen unix://. The reporter says the installed Desktop SSH bootstrap uses nohup for this command, which suggests an SSH bootstrap origin. The launch event was not retained, so this is a hypothesis rather than a confirmed cause (source: https://github.com/openai/codex/issues/52675). Reporter-applied recovery: codex app-server daemon stop said the server was not managed by the daemon manager. After confirming the background server had no loaded threads, the reporter sent SIGTERM only to it. The Desktop app-server then changed from Errored to Connected. No thread, database, or lock data was modified, and remote delivery had not yet been retested (source: https://github.com/openai/codex/issues/52675). Status and limitations: the issue is open with no maintainer comment. The only comment is an automated bot note flagging possible duplicate #52099 (source: https://github.com/openai/codex/issues/52675#issuecomment-6087402143). No fix commit or release is cited. This is a single external report; KFA did not reproduce it. The auto-update trigger is not established, and the background server's version is unknown. The reporter wants the single-writer safeguard to stay intact. Diagnosis is limited to checking for multiple app-server processes sharing one Codex home and which process holds the Remote Control connection and the writer lock.