feat(tools): texture generation through the documented Stability Matrix workflows

Re-read /home/zwitschi/StabilityMatrix/Data/README.md - it now documents 141 GB of models and four
verified, core-node image/texture graphs - and wired the ones that matter into the toolchain.

comfyui_generate.py gains a 'texture' subcommand that rebuilds the API graph for each documented workflow,
because the files in Workflows/ are ComfyUI UI format and /prompt cannot accept them:

  albedobase  AlbedoBase XL v3.1, 30 steps, CFG 6.0, dpmpp_2m/karras, baked VAE
  krea2       Krea-2 Turbo NVFP4, 8 steps, CFG 1.0, euler/simple, ConditioningZeroOut, texture-fix VAE
  qwen21      Qwen-Image 2.1 int8, 25 steps, CFG 1.0, euler/simple, TextEncodeQwenImage21, texture-fix VAE

Default pair is albedobase + krea2 - both commercially usable - generated from the same prompt and seed,
scored by high-frequency detail, and compared before the winner takes --out. Smoke-tested at 512 px: both
rendered, albedobase won 18.91 against 17.51, and the resulting limestone albedo is a plausible material
sheet.

Tools/README.md documents the profiles with their settings, a licence row per profile (Qwen-Image 2.1 is
research-only and must not ship), and the two limits the workflows carry: they produce albedo only, and
ComfyUI 0.44 has no seamless sampler, so tiling has to be checked by offsetting the image 50%. Five tests
cover the graph shapes; the suite is at 108.

Tools/README.md

@@ -85,12 +85,36 @@ Geographic tools live in [`Geo/tools/`](../Geo/README.md) and the Blender-side b
 
 | Tool | Purpose |
 | --- | --- |
-| `comfyui_generate.py` | Drives the **local** ComfyUI at `http://127.0.0.1:8188` (Stability Matrix install): `models` lists checkpoints, `txt2img` renders through the HTTP API and writes a provenance sidecar (model, prompt, seed, full graph). Two checkpoints by default — same prompt and seed, both scored by high-frequency detail, the winner takes `--out`, the loser is kept as `<out>__<model>.png`. `--alt-model ""` for a single run. Local and human-triggered; no credential. |
+| `comfyui_generate.py` | Drives the **local** ComfyUI at `http://127.0.0.1:8188` (Stability Matrix install). `models` lists checkpoints; `txt2img` renders through the HTTP API; **`texture` renders material sheets with the project's documented texture workflows** (see below). Every run writes a provenance sidecar (model, prompt, seed, settings, full graph). Two candidates by default — same prompt and seed, both scored by high-frequency detail, the winner takes `--out`, the loser is kept beside it. Local and human-triggered; no credential. |
 | `generate_facade_textures.py` | Procedural, deterministic facade texture maps (ashlar base colour + roughness, stucco, asphalt, concrete) written with a pure-zlib PNG encoder — no model, no network, byte-identical on every run. |
 | `generate_character_textures.py` | Procedural character texture maps, same deterministic approach as the facade set. |
 | `facade_detail_concepts.py` | Generates **frontal building plates** (whole elevation, entrance in its facade, or building with neighbours) from a local reference photograph with an image model (OpenRouter); close-up detail crops are deliberately not requested because they cannot be measured for proportion. `--list` is free and shows which studies have their reference present; `--all`/`--only <name>` generates. Never overwrites (writes `-v2`, `-v3`), writes a `.refgen.json` sidecar per image (model, prompt, source photo, SHA-256), and puts everything under `Art/reference/facade-details/` — gitignored, local-only, never shipped. Costs money per image, so it is never automatic. |
 | `judge_screenshot.py` | (see Visual review) |
 
+### Texture generation workflows (from the Stability Matrix install)
+
+The install documents four verified, core-node graphs in [`Workflows/texture_generation.md`](/home/zwitschi/StabilityMatrix/Data/Workflows/texture_generation.md). Those files are ComfyUI **UI format**, which `/prompt` cannot accept, so `comfyui_generate.py texture` rebuilds the equivalent API graphs with exactly their settings:
+
+| Profile | Model | Steps / CFG / sampler | VAE | Best for | Licence |
+| --- | --- | --- | --- | --- | --- |
+| `albedobase` | `albedobaseXL_v31Large.safetensors` | 30 / 6.0 / `dpmpp_2m` `karras` | baked | stone, brick, metal, wood albedo | Civitai — commercial, credit required |
+| `krea2` | `krea2_turbo_nvfp4.safetensors` | 8 / 1.0 / `euler` `simple` | texture-fix (Qwen 1 / Krea-2) | fast stochastic detail: stone, bark, foliage, fabric | Krea-2 community licence |
+| `qwen21` | `qwen_image_2.1_int8_convrot.safetensors` | 25 / 1.0 / `euler` `simple` | texture-fix (Qwen 2.1) | material sheets with readable labels/signage, 2K | **Qwen research — non-commercial** |
+
+```bash
+python3 Tools/comfyui_generate.py texture \
+    --material "weathered grey limestone ashlar, fine grain" \
+    --out Art/textures/generated/T_StoneAshlar_AI.png
+```
+
+The default pair is `albedobase` + `krea2` — both commercially usable — generated from one prompt and seed, scored, and compared before the winner takes `--out`. `--profile`/`--alt-profile` pick another pair, and `--alt-profile ""` runs one.
+
+**Limits this tool inherits from the workflows, and does not hide:**
+
+- The graphs produce a **colour/albedo sheet only**. Normal, roughness and metallic maps are derived downstream with a material tool.
+- ComfyUI 0.44 core has **no seamless-tiling sampler**, so "seamless tileable" in the prompt is a hint, not a guarantee. Offset the image 50 % and check for a seam before using it on a tiled surface; the installed SDXL tile ControlNet is the fix-up path.
+- A generated texture ships only once its model and licence are in [`THIRD_PARTY_NOTICES.md`](../THIRD_PARTY_NOTICES.md).
+
 ### Local image generation: start Stability Matrix first
 
 `comfyui_generate.py` talks to ComfyUI on `http://127.0.0.1:8188`. That server is **not** started by this repository — it lives in a [Stability Matrix](https://lykos.ai) install and must already be running.