diff --git a/client/tests/visual_capture.gd b/client/tests/visual_capture.gd
index 8bcb08e88..b54c3a31b 100644
--- a/client/tests/visual_capture.gd
+++ b/client/tests/visual_capture.gd
@@ -696,6 +696,11 @@ func _print_list() -> void:
 	var flows: Dictionary = _config.get("flows", {})
 	print("scenarios:")
 	for name in scenarios:
+		# Skip "_"-prefixed comment keys and any non-Dictionary value — a bare
+		# string here crashed --list once (PR #180: a provenance _comment key
+		# placed inside scenarios{}; now top-level, but stay robust).
+		if String(name).begins_with("_") or not (scenarios[name] is Dictionary):
+			continue
 		var desc: String = scenarios[name].get("description", "")
 		print("  %s — %s" % [name, desc])
 	# T-1120: atlas_matrix isn't a tests/visual.json scenarios{} entry (it reads
diff --git a/tests/visual.json b/tests/visual.json
index 58303bc9d..2a6168464 100644
--- a/tests/visual.json
+++ b/tests/visual.json
@@ -4,6 +4,7 @@
   "tolerance": 5,
   "max_diff_pct": 0.5,
   "golden_dir": "client/tests/golden/visual",
+  "_comment_atlas_goldens": "T-1120 provenance (PR #180 review, Tyre): the 12 atlas_* golden scenarios were captured AND verified (exact pixel match) on danoontje's Mesa/AMD stack. T-1121 documents that goldens do NOT port across boxes at the current tolerance (the 19 pre-existing goldens fail 12-14% on this same box) — an atlas_* golden failing on a DIFFERENT machine is expected baseline drift, not a regression, until T-1121 resolves the canonical-capture-box question.",
   "scenarios": {
     "fog_3state": {
       "ticks": 5,
@@ -90,7 +91,6 @@
       "env": { "SR_AUTOPILOT": "stance_up,south:6.0" },
       "description": "Live 3D sandbox: autopilot walks south into a wall — cutaway stub golden (T-1088)"
     },
-    "_comment_atlas_goldens": "T-1120 provenance (PR #180 review, Tyre): the 12 atlas_* goldens below were captured AND verified (exact pixel match) on danoontje's Mesa/AMD stack. T-1121 documents that goldens do NOT port across boxes at the current tolerance (the 19 pre-existing goldens fail 12-14% on this same box) — an atlas_* golden failing on a DIFFERENT machine is expected baseline drift, not a regression, until T-1121 resolves the canonical-capture-box question.",
     "atlas_GJ380c_zfit_C": {
       "ticks": 240,
       "live": true,