The agent's comment on socket removal describes how a flan dev session removes it now.

vendor/agent/flan_agent.c

@@ -1035,15 +1035,10 @@ static int sock_dir_fd = -1;
  * Outside a daemon that is the whole story, and the path is under TMPDIR where
  * nothing else would ever reclaim it. SIGKILL is the one end nothing survives.
  *
- * Under [flan dev] none of the three is how a session usually ends, and this
- * is worth being exact about rather than claiming cover it does not give: the
- * two-process daemon kills its child with SIGTERM, whose default disposition
- * runs no atexit, and the merged session leaves by [Unix._exit 0]. So the
- * socket there is left in the daemon's temp directory — which is itself never
- * removed today. TODO.org, "The daemon leaves its temp directory behind": the
- * session-end cleanup that item asks for takes the socket with it, and until
- * it lands the socket outlives the session. Nothing below can fix that from
- * here; a program that is killed does not get to tidy up. */
+ * Under [flan dev] the session's own end also covers it: the two-process
+ * daemon stops its child with SIGTERM, which [fatal_unlink] sees, and the
+ * merged session leaves by [Unix._exit 0] after removing its temp directory,
+ * the socket with it (lib/dev.ml, [remove_session_dirs]). */
 static void unlink_bound_sock(void) {
   if (bound_sock[0] != '\0') unlink(bound_sock);
 }