Pluviobyte · Pluviobyte · Jul 17, 2026 · Jul 17, 2026
diff --git a/.claude-plugin/marketplace.json b/.claude-plugin/marketplace.json
@@ -1,8 +1,8 @@
 {
   "name": "rnskill",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "owner": "Pluviobyte",
-  "description": "雪踏乌云的 AI Agent Skills 集合 — 写作精修 + 动效制作 + 参考复刻 + 全帧质检",
+  "description": "雪踏乌云的 AI Agent Skills 集合 — 写作精修 + 内容提取 + 图像生成 + 动效制作 + 参考复刻 + 全帧质检",
   "repository": "https://github.com/Pluviobyte/rnskill",
   "license": "CC-BY-NC-4.0",
   "plugins": [
@@ -26,6 +26,16 @@
       "keywords": ["视频", "动效", "导演", "motion", "director", "video"],
       "skills": ["./skills/rn-motion-director"]
     },
+    {
+      "name": "rn-ian-xiaohei-cat-illustrations",
+      "description": "Ian 小黑猫怪诞正文配图。把中文文章中的判断、流程和隐喻转成 16:9 白底手绘插图，固定使用小黑猫 IP。改编自 Ian 的原版 Xiaohei Skill。\nIan-style Xiaohei Cat article illustrations with sparse hand-drawn 16:9 compositions and recurring character consistency.\nTrigger: /rn-ian-xiaohei-cat-illustrations, 小黑猫, 黑猫配图, 正文配图, 文章插图, 怪诞手绘, shot list",
+      "source": "./",
+      "strict": false,
+      "version": "1.0.0",
+      "category": "image-generation",
+      "keywords": ["小黑猫", "正文配图", "文章插图", "手绘", "shot list", "xiaohei", "illustration"],
+      "skills": ["./skills/rn-ian-xiaohei-cat-illustrations"]
+    },
     {
       "name": "rn-motion-replica",
       "description": "把获授权的参考片段拆成视觉证据，并制作成原创、可编辑、可验证的 HyperFrames 动效工程。\nOriginal editable motion study from an authorized reference range, with analysis evidence and final-MP4 QC.\nTrigger: /rn-motion-replica, 参考这个视频做同款动效, 复刻前N秒方框连线动画",
@@ -36,6 +46,16 @@
       "keywords": ["参考视频", "动效复刻", "HyperFrames", "motion", "replica", "editable"],
       "skills": ["./skills/rn-motion-replica"]
     },
+    {
+      "name": "rn-editorial-collage-motion",
+      "description": "半色调纸张拼贴动效。参考拆解 → 可编辑 JSON 规范 → Codex 静帧审批 → 本地逐层组装动画，无需 Arcads MCP。\nHalftone paper-collage motion with Codex-generated stills and local assemble-from-empty rendering. No Arcads MCP required.\nTrigger: /rn-editorial-collage-motion, 半色调拼贴, 纸张拼贴, 拼贴动效, 从无到有, editorial collage, halftone collage, animate this",
+      "source": "./",
+      "strict": false,
+      "version": "1.0.0",
+      "category": "video-production",
+      "keywords": ["半色调拼贴", "纸张拼贴", "拼贴动效", "从无到有", "halftone", "collage", "motion"],
+      "skills": ["./skills/rn-editorial-collage-motion"]
+    },
     {
       "name": "rn-replica-qc",
       "description": "参考视频复刻 SOP v2。五级保真度 + 素材/运行时/交付三道全帧门 + 精确重放与参数化组件分离。\nReference replica SOP v2 with five fidelity levels and asset/runtime/delivery frame gates.\nTrigger: /rn-replica-qc, 复刻这个视频, 对齐参考视频, 逐帧像素验证",

diff --git a/README.md b/README.md
@@ -4,7 +4,7 @@
 
 AI Agent Skills maintained by 雪踏乌云 for Codex, Claude Code, and other Agent workflows that support `SKILL.md`.
 
-The collection covers Chinese writing, content extraction, motion direction, original reference-motion studies, style-specific video production, openers, and evidence-based replica QA.
+The collection covers Chinese writing, content extraction, recurring-character illustration generation, motion direction, original reference-motion studies, halftone collage motion, style-specific video production, openers, and evidence-based replica QA.
 
 ## Requirements
 
@@ -29,7 +29,7 @@ npx -y skills add Pluviobyte/rnskill -g --all
 Or install a single skill:
 
 ```bash
-npx -y skills add Pluviobyte/rnskill --skill rn-renhua
+npx -y skills add Pluviobyte/rnskill --skill rn-editorial-collage-motion
 ```
 
 ### Manual Install
@@ -41,11 +41,15 @@ Copy only the skill you need into your project:
 mkdir -p <project>/.agents/skills
 cp -R skills/rn-renhua <project>/.agents/skills/rn-renhua
 cp -R skills/rn-motion-replica <project>/.agents/skills/rn-motion-replica
+cp -R skills/rn-editorial-collage-motion <project>/.agents/skills/rn-editorial-collage-motion
+cp -R skills/rn-ian-xiaohei-cat-illustrations <project>/.agents/skills/rn-ian-xiaohei-cat-illustrations
 
 # Claude Code
 mkdir -p <project>/.claude/skills
 cp -R skills/rn-renhua <project>/.claude/skills/rn-renhua
 cp -R skills/rn-motion-replica <project>/.claude/skills/rn-motion-replica
+cp -R skills/rn-editorial-collage-motion <project>/.claude/skills/rn-editorial-collage-motion
+cp -R skills/rn-ian-xiaohei-cat-illustrations <project>/.claude/skills/rn-ian-xiaohei-cat-illustrations
 ```
 
 ## Available Skills
@@ -62,12 +66,19 @@ cp -R skills/rn-motion-replica <project>/.claude/skills/rn-motion-replica
 |-------|-------------|
 | [`rn-wechat-extract`](skills/rn-wechat-extract/) | Extract full text from WeChat public account articles via MicroMessenger UA spoofing. Stdlib only, no API key. |
 
+### Image Generation
+
+| Skill | Description |
+|-------|-------------|
+| [`rn-ian-xiaohei-cat-illustrations`](skills/rn-ian-xiaohei-cat-illustrations/) | Generates sparse 16:9 Chinese article illustrations with a recurring deadpan Xiaohei Cat character, white hand-drawn compositions, short colored annotations, and explicit identity-consistency QA. |
+
 ### Video Production
 
 | Skill | Description |
 |-------|-------------|
 | [`rn-motion-director`](skills/rn-motion-director/) | Motion-first AI video director. Turns topics into motion video concepts with visual metaphors, beat graphs, and anti-PPT QC. |
 | [`rn-motion-replica`](skills/rn-motion-replica/) | Builds an original, editable HyperFrames motion study from an authorized reference range, with analysis evidence and final-MP4 QC. |
+| [`rn-editorial-collage-motion`](skills/rn-editorial-collage-motion/) | Turns a reference or short brief into an editable halftone paper-collage spec, approved Codex-generated stills, and deterministic assemble-from-empty motion rendered locally with FFmpeg or HyperFrames. |
 | [`rn-dark-saas-video`](skills/rn-dark-saas-video/) | Dark cinematic SaaS product video in "magic UI" style. 8 scene blueprints, 3 timing presets, hard style rules. |
 | [`rn-bw-text-opener`](skills/rn-bw-text-opener/) | Black-white typed text opener animation with synced typing SFX. 3 timing presets. Includes a Python timing plan generator. |
 
@@ -86,6 +97,8 @@ rnskill/
 │   ├── rn-wechat-extract/      # Extraction: WeChat article reader
 │   ├── rn-motion-director/     # Video: motion director
 │   ├── rn-motion-replica/      # Video: original editable motion study
+│   ├── rn-editorial-collage-motion/ # Video: halftone collage assembly
+│   ├── rn-ian-xiaohei-cat-illustrations/ # Image: Xiaohei Cat article illustrations
 │   ├── rn-dark-saas-video/     # Video: dark SaaS style
 │   ├── rn-bw-text-opener/      # Video: typed text opener
 │   └── rn-replica-qc/          # QC: reference video replica
@@ -105,9 +118,21 @@ python3 tools/sync-video-skills.py --source /path/to/video-production-skills
 python3 tools/sync-video-skills.py --source /path/to/video-production-skills --check
 ```
 
+## Credits and Adaptations
+
+### Xiaohei Cat Illustrations
+
+`rn-ian-xiaohei-cat-illustrations` is derived from Ian's original [Ian Xiaohei Illustrations](https://github.com/helloianneo/ian-xiaohei-illustrations), referred to here as the original Xiaohei Skill. It preserves the article-analysis, shot-list, white-background hand-drawn visual system, composition patterns, and calibration examples, then adapts the recurring black figure into a locally defined Xiaohei Cat IP with authoritative character references and stricter identity-consistency QA. The upstream project is MIT-licensed; its full copyright and license notice is retained in [`origin-and-license.md`](skills/rn-ian-xiaohei-cat-illustrations/references/origin-and-license.md). This adaptation is not an official Ian release.
+
+### Collage Motion
+
+`rn-editorial-collage-motion` is an independent adaptation inspired by Vikash Kumar's original [Arcads Collage Motion Skill](https://buldrr.com/arcads-collage-motion-skill/). Credit goes to Vikash Kumar for the original two-stage idea: decode a reference into an editable visual specification, then animate approved collage stills as an assemble-from-empty sequence.
+
+The original workflow requires the Arcads MCP connector and Arcads credits, and uses Nano Banana 2 for stills plus Seedance 2.0 for motion. This repository edition replaces that service chain with Codex's built-in image generation and local FFmpeg or HyperFrames rendering. It therefore needs no Arcads MCP connection and consumes no Arcads credits, making it a free local-compatible edition for users who already have access to Codex; existing Codex access and local compute requirements still apply. This adaptation is not an official Arcads or original-author release.
+
 ## License
 
-CC BY-NC 4.0. See [LICENSE](LICENSE).
+Unless otherwise noted, CC BY-NC 4.0. See [LICENSE](LICENSE). Adapted third-party components retain their upstream notices; the Xiaohei-derived files include Ian's MIT notice in [`origin-and-license.md`](skills/rn-ian-xiaohei-cat-illustrations/references/origin-and-license.md).
 
 ## Author
 

diff --git a/README.zh.md b/README.zh.md
@@ -4,7 +4,7 @@
 
 雪踏乌云维护的 AI Agent Skills 集合，适用于 Codex、Claude Code 等支持 `SKILL.md` 的 Agent 工作流。
 
-当前覆盖中文写作、内容提取、动效导演、原创参考动效复刻、风格化视频、片头包装和带逐帧证据的复刻质检。
+当前覆盖中文写作、内容提取、固定角色插画生成、动效导演、原创参考动效复刻、半色调拼贴动效、风格化视频、片头包装和带逐帧证据的复刻质检。
 
 ## 前置要求
 
@@ -29,7 +29,7 @@ npx -y skills add Pluviobyte/rnskill -g --all
 安装单个 Skill：
 
 ```bash
-npx -y skills add Pluviobyte/rnskill --skill rn-renhua
+npx -y skills add Pluviobyte/rnskill --skill rn-editorial-collage-motion
 ```
 
 ### 手动安装
@@ -41,11 +41,15 @@ npx -y skills add Pluviobyte/rnskill --skill rn-renhua
 mkdir -p <project>/.agents/skills
 cp -R skills/rn-renhua <project>/.agents/skills/rn-renhua
 cp -R skills/rn-motion-replica <project>/.agents/skills/rn-motion-replica
+cp -R skills/rn-editorial-collage-motion <project>/.agents/skills/rn-editorial-collage-motion
+cp -R skills/rn-ian-xiaohei-cat-illustrations <project>/.agents/skills/rn-ian-xiaohei-cat-illustrations
 
 # Claude Code
 mkdir -p <project>/.claude/skills
 cp -R skills/rn-renhua <project>/.claude/skills/rn-renhua
 cp -R skills/rn-motion-replica <project>/.claude/skills/rn-motion-replica
+cp -R skills/rn-editorial-collage-motion <project>/.claude/skills/rn-editorial-collage-motion
+cp -R skills/rn-ian-xiaohei-cat-illustrations <project>/.claude/skills/rn-ian-xiaohei-cat-illustrations
 ```
 
 ## 可用技能
@@ -62,12 +66,19 @@ cp -R skills/rn-motion-replica <project>/.claude/skills/rn-motion-replica
 |-------|------|
 | [`rn-wechat-extract`](skills/rn-wechat-extract/) | 微信公众号文章全文提取。通过 MicroMessenger UA 伪装绕过微信访问控制，纯标准库，无需 API key。 |
 
+### 图像生成
+
+| Skill | 说明 |
+|-------|------|
+| [`rn-ian-xiaohei-cat-illustrations`](skills/rn-ian-xiaohei-cat-illustrations/) | 为中文文章生成 16:9 白底手绘正文配图，以死认真、略怪诞的小黑猫为固定角色，并通过权威参考图和角色一致性评分保持 IP 稳定。 |
+
 ### 视频制作
 
 | Skill | 说明 |
 |-------|------|
 | [`rn-motion-director`](skills/rn-motion-director/) | AI 动效导演元 Skill。把选题/脚本转化为动效视频概念：视觉隐喻、运动语法、场景节拍图、Anti-PPT 质量门。 |
 | [`rn-motion-replica`](skills/rn-motion-replica/) | 从获授权的参考片段构建原创、可编辑的 HyperFrames 动效工程，附参考分析证据和最终 MP4 质检。 |
+| [`rn-editorial-collage-motion`](skills/rn-editorial-collage-motion/) | 将参考图或简述拆成可编辑的半色调纸张拼贴规范，使用 Codex 生成并审批静帧，再用本地 FFmpeg 或 HyperFrames 制作“从空背景逐件组装”的确定性动画。 |
 | [`rn-dark-saas-video`](skills/rn-dark-saas-video/) | 暗色 SaaS 魔术短片。黑色星空舞台 + 紫色底光 + 大字动效 + 渐变 CTA。8 套场景蓝图、3 种时长预设。 |
 | [`rn-bw-text-opener`](skills/rn-bw-text-opener/) | 黑白文字打字机开场动画。纯黑背景 + 白色逐字打字 + 同步音效 + 文字替换效果。3 种时长预设，附带 Python 时序规划脚本。 |
 
@@ -86,6 +97,8 @@ rnskill/
 │   ├── rn-wechat-extract/      # 提取：微信公众号文章
 │   ├── rn-motion-director/     # 视频：动效导演
 │   ├── rn-motion-replica/      # 视频：原创可编辑动效复刻
+│   ├── rn-editorial-collage-motion/ # 视频：半色调拼贴组装动效
+│   ├── rn-ian-xiaohei-cat-illustrations/ # 图像：小黑猫正文配图
 │   ├── rn-dark-saas-video/     # 视频：暗色 SaaS 风格
 │   ├── rn-bw-text-opener/      # 视频：黑白打字开场
 │   └── rn-replica-qc/          # 质检：参考视频复刻
@@ -105,9 +118,21 @@ python3 tools/sync-video-skills.py --source /path/to/video-production-skills
 python3 tools/sync-video-skills.py --source /path/to/video-production-skills --check
 ```
 
+## 致谢与改编说明
+
+### 小黑猫正文配图
+
+`rn-ian-xiaohei-cat-illustrations` 来源于 Ian 的原版 [Ian Xiaohei Illustrations](https://github.com/helloianneo/ian-xiaohei-illustrations)，也就是原始 Xiaohei Skill（小黑 Skill）。当前版本保留了文章认知锚点分析、shot list、16:9 白底手绘视觉系统、构图方法和校准样例，并将原版黑色“小黑”角色进一步改造成固定“小黑猫”IP，增加权威角色参考图和更严格的身份一致性质检。原项目采用 MIT 许可，完整版权与许可证保留在 [`origin-and-license.md`](skills/rn-ian-xiaohei-cat-illustrations/references/origin-and-license.md)。本改编版不是 Ian 发布的官方版本。
+
+### 半色调拼贴动效
+
+`rn-editorial-collage-motion` 是基于 Vikash Kumar 原版 [Arcads Collage Motion Skill](https://buldrr.com/arcads-collage-motion-skill/) 思路制作的独立改编版。感谢 Vikash Kumar 提出的两阶段工作流：先把参考视觉解码为可编辑规范，再把获批的拼贴静帧做成“从无到有”的逐件组装动画。
+
+原版工作流需要连接 Arcads MCP 并消耗 Arcads credits，静帧和视频分别使用 Nano Banana 2 与 Seedance 2.0。本仓库版本保留参考拆解、可编辑 JSON 规范、静帧审批和逐层组装逻辑，但将执行链路优化为 Codex 内置生图，加本地 FFmpeg 或 HyperFrames 渲染。因此它不需要 Arcads MCP，也不消耗 Arcads credits；对于已经可以使用 Codex 的用户，这是一个本地免费兼容版，但仍受用户自身 Codex 使用权限和本机算力条件影响。本改编版不是 Arcads 或原作者发布的官方版本。
+
 ## 许可证
 
-CC BY-NC 4.0。详见 [LICENSE](LICENSE)。
+除另有说明外，仓库采用 CC BY-NC 4.0，详见 [LICENSE](LICENSE)。改编自第三方的内容继续保留其上游许可；小黑 Skill 衍生文件中的 Ian MIT 许可见 [`origin-and-license.md`](skills/rn-ian-xiaohei-cat-illustrations/references/origin-and-license.md)。
 
 ## 作者
 

diff --git a/VERSION b/VERSION
@@ -1 +1 @@
-1.1.0
+1.2.0