@goodandready/dsh-image-gen 0.10.32 → 0.10.33

package/README.md

@@ -123,6 +123,11 @@
 ## 🛠️ Complete Tools Reference
 
 * **`generate_image`**: Generate images with pluggable providers, seeds, aspect ratios, and style presets.
+* **`set_style_anchor` (#283)**: Lock session visual identity (subject, outfit, medium, color palette, lighting) for consistent multi-turn generation.
+* **`generate_ui_asset` (#284)**: Layout-aware UI asset generator for icons, stickers, badges, spot graphics, and heroes with negative space constraints (`isolated`, `centered`, `left_empty`, `right_empty`, `top_empty`, `bottom_empty`) and auto-transparency.
+* **`generate_style_matrix` (#286)**: 2x2 visual benchmarking matrix exploring 4 aesthetic presets simultaneously with blind compare mode and 1-click style preset adoption.
+* **`In-Chat Inpainting Canvas` (#285)**: Interactive brush overlay and mask drawer mounted directly over image message cards in chat.
+* **`Smart Provider Fallback Chain` (#282)**: Intelligent cascading failover across providers on 429, 5xx, timeouts, or depleted balance.
 * **`edit_image`**: Targeted inpainting and modification with automatic session reference resolution (#142, #144, #145).
 * **`vary_image`**: Controlled variation generation preserving composition (#143, #144).
 * **`remove_background`**: Extract subject with transparent PNG output (FAL BiRefNet / Rembg).

package/README.ru.md

@@ -163,6 +163,9 @@ graph LR
 | Инструмент | Назначение | Ключевые параметры |
 |---|---|---|
 | **`generate_image`** | Генерация изображения по тексту | `prompt`, `image_size`, `seed`, `style`, `negative_prompt`, `count` |
+| **`set_style_anchor`** | Фиксация визуального стиля и персонажа сессии (объект, костюм, палитра, свет) (#283) | `mode` (`character`/`style`), `image`, `description`, `strength`, `clear` |
+| **`generate_ui_asset`** | Генерация интерфейсных ассетов с контролем свободного пространства и авто-прозрачностью (#284) | `prompt`, `asset_type`, `layout_composition`, `transparent`, `output_format` |
+| **`generate_style_matrix`** | Сетка 2x2 для одновременного бенчмаркинга концепта в 4 стилях со слепым сравнением (#286) | `prompt`, `styles`, `blind_mode`, `output_format` |
 | **`edit_image`** | Направленное редактирование / inpainting с авто-поиском оригинала (#142, #144, #145) | `prompt`, `image` (по умолч. `latest`), `mask`, `strength` |
 | **`vary_image`** | Генерация вариаций существующей картинки (#143, #144) | `image` (по умолч. `latest`), `prompt`, `variation_strength` (0.1–0.9), `count` |
 | **`remove_background`** | Удаление фона с сохранением прозрачного PNG | `image`, `model`, `output_name` |

package/README.zh.md

@@ -113,4 +113,12 @@ dsh plugin --profile web add @goodandready/dsh-image-gen
 
 ## 📄 许可证
 
-MIT © [GooDAnDReaDY](https://github.com/GooDAnDReaDY)
+MIT © [GooDAnDReaDY](https://github.com/GooDAnDReaDY)
+
+## 🛠️ 新增核心功能 (v0.10.32)
+
+* **`set_style_anchor` (#283)**: 会话级角色与风格参考锚点，在连续对话中锁定主体、配色、光照与渲染媒介。
+* **`generate_ui_asset` (#284)**: 布局感知与留白控制 UI 资产生成器，支持精准留白（`isolated`、`centered`、`left_empty`、`right_empty`、`top_empty`、`bottom_empty`）与自动透明通道。
+* **`generate_style_matrix` (#286)**: 2x2 风格矩阵探索与双盲对比，并发生成 4 种美学预设切片并支持一键应用。
+* **`对话内局部重绘画布` (#285)**: 在消息卡片上直接唤起交互式笔刷画布，支持选区涂抹与即时 Inpainting 局部重绘。
+* **`智能服务商自动故障转移链` (#282)**: 主服务商 429、5xx、超时或额度耗尽时自动无缝切换备选渠道。

package/lib/tools/style-matrix.js

@@ -94,7 +94,7 @@ export function registerStyleMatrixTools(ctx, deps) {
                     seed: { type: 'integer' },
                     width: { type: 'integer' },
                     height: { type: 'integer' },
-                    attachment: { type: 'object' },
+                    attachment: { type: 'object', additionalProperties: true },
                   },
                 },
               },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@goodandready/dsh-image-gen",
-  "version": "0.10.32",
+  "version": "0.10.33",
   "description": "Image generation for DeepSeek Harness: a generate_image tool with pluggable providers — the FAL queue, any OpenAI-compatible images API, or a ChatGPT/Grok subscription with no API key at all. The picture is shown inline in the conversation; the model receives either a link (works with any chat model) or the image itself (needs dsh-vision-bridge or a vision-capable model).",
   "keywords": [
     "deepseek-harness",