orchestrator: emit snapshot_load before _write_meta to keep t_mono ~0

On slower disks (Pi5 SD cards, mu's hardware) the json.dump → write →
os.replace path inside _write_meta takes more than 1 ms, so when the
snapshot_load event fired afterwards its t_mono_ns drifted past the
"<1 ms after origin" assertion in test_driver_events_persist_to_events_jsonl.

Fix: emit snapshot_load immediately after setting _t_mono_origin_ns,
before any file I/O. Matches the semantic intent (snapshot_load marks
episode clock = 0) and removes the disk-speed dependency from the
event timeline.

Diagnosis + suggested patch from spectral/CIS490#7 (filed by mu).

Closes spectral/CIS490#7.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

orchestrator/episode.py

@@ -124,13 +124,15 @@ class EpisodeRunner:
 
     def run(self) -> EpisodeResult:
         self._t_mono_origin_ns = time.monotonic_ns()
-        started_at_wall = datetime.now(timezone.utc).isoformat()
+        # snapshot_load is the marker for "episode clock = 0". Emit
+        # BEFORE any file I/O — _write_meta() takes >1 ms on slow disks
+        # (Refs spectral/CIS490#7).
+        self.emit_event("snapshot_load", snapshot=self.cfg.snapshot_name)
 
+        started_at_wall = datetime.now(timezone.utc).isoformat()
         meta = self._initial_meta(started_at_wall)
         self._write_meta(meta)
 
-        self.emit_event("snapshot_load", snapshot=self.cfg.snapshot_name)
-
         # Snapshot revert at start: pause+restore the guest to a known
         # baseline before phase 0. Requires QMP and a savevm having
         # already taken place (the launcher is responsible for that).