test(review): judge every shot from the targets file, and record that all nine fail

--shots was parsing a table shape the README no longer has, so it silently found nothing.
It now reads Tools/visual_targets.json, the same file the render gate uses, and reports one
line per shot with the model's verdict and reasoning.

The result is not good news and is committed as such: 0 of 9 shots judged 'yes'. Shot 7 shows
the scene itself wrong - a disassembled car floating over a blank surface with no street or
buildings in frame - and the others fail for missing detail rather than for a broken frame.
visual_loop --verify still reports 9/9, because it only measures exposure and size.

A warning now sits at the top of Screenshots/README.md: treat every image there as stale until
the scene is fixed. The cause is not yet identified - the map builder's recent changes are 15
lines of vehicle staging, which does not explain a missing street.

Screenshots/README.md

@@ -12,6 +12,8 @@ Tools/render_screenshots.sh --verify
 
 Stages the demo shot cameras, runs the game on `/Game/Hollywood/DemoMap`, copies the PNGs into `Screenshots/demo/` and then runs [`Tools/visual_loop.py`](../Tools/visual_loop.py) `--verify` to check that each shot still shows the content its expectation describes. About a minute.
 
+> **WARNING — the current capture set is broken (2026-11-09).** All nine shots were judged `no` by the vision model, and shot 7 shows the scene itself wrong: a disassembled car floating over a blank surface with no street or buildings in frame. `--verify` still reports 9/9 because it only measures exposure and size. Treat every image in this file as stale until this is fixed; the cause is not yet identified.
+
 ## Status at a glance
 
 Every shot is embedded below, in order, so the whole review happens in this file.