[lane D] The last two pathogens: the seeker and the turret get meshes

Completes the set. These two matter most of the five because they are the most COMMON enemies in
the game — the seeker is pepsin, eosinophil and neutrophil across four levels, and the turret is
every candida bloom and Peyer's patch.

  seeker -> amoeboid immune cell, rounded body with blunt pseudopod lobes. 3194 tris, replaces a
            6-sided cone.
  turret -> candida plaque colony, a lumpy dome of fused budding yeast. 3894 tris, replaces a
            7-sided cylinder.

Both took two MODELBEAST runs: the farm 401'd then 429'd the guest token twice under load, which
is a rate limit rather than a content failure, and the script being resumable meant the retry
only paid for what was missing.

A judgement call worth recording: the seeker's bbox aspect is 1.16, near-round, which for the
phage was the signature of a FAILED reconstruction. Here it is correct — an amoeboid cell is
round, and the detail is in the lobes rather than the silhouette. The cost is that a charging
seeker no longer has the cone's "pointy end = coming at you" read. The fiction wins: pepsin and
eosinophils are cells, not missiles, and the thing is 1 unit across at 19 u/s, so its threat is
read from motion and colour long before shape.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

web/assets/manifest.json

@@ -72,6 +72,14 @@
     "driller": {
       "url": "models/driller.glb",
       "tris": 3231
+    },
+    "seeker": {
+      "url": "models/seeker.glb",
+      "tris": 3194
+    },
+    "turret": {
+      "url": "models/turret.glb",
+      "tris": 3894
     }
   },
   "textures": {