Add OpenRouter AI texture generation mode to the asset pipeline

- tools/generate_textures.py: --ai mode (OPENROUTER_API_KEY, --prompt,
  --name, --model, --size; base64 PNG decode; clamp to <=1024px) with a
  default batch covering the authored texture set; seedream minimum-size
  handled by requesting 2048 and downscaling
- Generated AI asset set committed under assets/textures/ai/ (15 files,
  1024x1024 each), kept separate from the deterministic PIL set
- README/MEMORY/TODO document the mode and live verification
- (work authored in a parallel session; linted/verified here — 57/57 tests)

MEMORY.md

@@ -59,6 +59,14 @@ Session memory for durable decisions, state, and gotchas. Keep concise, remove o
 - Assets pipeline: `python3 tools/generate_textures.py` produces all authored PNGs under
   `assets/textures/` (≤1024px, pow2); regenerate after changing the script, then run
   `godot --headless --import` before tests. 57 headless tests guard the pipeline.
+- AI texture mode: `python3 tools/generate_textures.py --ai` generates textures via the
+  OpenRouter image API (`POST https://openrouter.ai/api/v1/images`, model
+  `bytedance-seed/seedream-4.5` by default, base64 PNG in `data[].b64_json`). Requires
+  `OPENROUTER_API_KEY` (a gitignored `.env` in the repo root holds it locally; source it
+  with `set -a && . ./.env && set +a`). `--prompt`/`--name`/`--model`/`--size` override the
+  default batch. Output is downscaled to ≤1024px. NOTE: seedream rejects sizes below
+  3,686,400 px (1920×1920), so the script requests `GEN_SIZE=2048` and downscales — do not
+  lower `GEN_SIZE` below 1920. Verified end-to-end 2026-10-31 (3 textures generated).
 - Detailed graphics: textured facade/awning/skyline, humanoid NPC rigs + props + walk bob +
   attention head-turn, inspection artifact cards, mode icons, escalation vignette, body-cam
   overlay (REC + shift clock). Screenshot set regenerated (`docs/screenshots/`, day + night).