Expand functional visual recipe coverage

The community agent needs more retrievable visual material, not just more showcase cases, so this adds 100 debranded functional visual recipes sourced from existing cloned prompt libraries and updates the public asset targets to 240 recipes.

Constraint: Public repo must not ship upstream prompt_sources paths or private image assets
Rejected: Add only gallery display cases | the user clarified the request was functional recipe expansion
Confidence: high
Scope-risk: moderate
Directive: Keep visual-recipes.json as retrievable agent input; do not treat these entries as passive gallery metadata
Tested: pnpm --filter ad-image-agent-core test
Tested: pnpm --filter ad-image-agent-core build
Tested: pnpm --filter ad-image-agent typecheck
Tested: pnpm --filter ad-image-agent build
Tested: pnpm --dir community-export --filter ad-image-agent-core test
Tested: pnpm --dir community-export --filter ad-image-agent-core build
Tested: pnpm --dir community-export --filter ad-image-agent typecheck
Tested: pnpm --dir community-export --filter ad-image-agent build
Tested: node scripts/validate-community.mjs community-export
Not-tested: Live Image2 generation quality for the new recipes

Author: kwistzz

README.md

@@ -4,7 +4,7 @@ Advertising prompt agent for Image2: turn business briefs and reference images i
 
 This is not another prompt dump. Prompt libraries are useful for browsing examples; Image2 Ads Studio is a working prompt pipeline for advertising production. It parses a customer brief, retrieves advertising templates and visual recipes, rewrites references through an LLM prompt brain, validates deterministic execution details, and returns an optimized prompt that can be manually tested in Image2.
 
-This repository is the **Community Edition**. It opens the agent framework, local Web workbench, all current 220 business templates, 140 visual recipes, LLM prompt brain interface, deterministic prompt checks, reference image policy, and a public case gallery.
+This repository is the **Community Edition**. It opens the agent framework, local Web workbench, all current 220 business templates, 240 visual recipes, LLM prompt brain interface, deterministic prompt checks, reference image policy, and a public case gallery.
 
 ![Workflow](docs/assets/github-workflow.svg)
 
@@ -51,7 +51,7 @@ Image2 Ads Studio turns that loose brief into a structured advertising plan and
 ## Core Features
 
 - Advertising vertical coverage: storefront signboards, posters, roll-up banners, event backdrops, wayfinding signage, brand walls, local promotions, ecommerce main images, product ads, and commercial photography.
-- Open template library: 220 business templates and 140 visual recipes.
+- Open template library: 220 business templates and 240 visual recipes.
 - LLM prompt brain: uses a Responses-compatible LLM endpoint to refine the rule prompt.
 - Reference image policy: user uploads can be read by the LLM and marked for later Image2 editing; upstream references are only used for visual understanding.
 - Deterministic prompt validation: removes vague expressions and requires composition, lighting, typography, material, and image-reference policy.
@@ -97,7 +97,7 @@ flowchart LR
 | Local Web UI | Included | Included |
 | Core prompt agent framework | Included | Included |
 | Business templates | 220 open templates | Custom/private templates and governance |
-| Visual recipes | 140 open recipes | Custom/private recipes and scored evaluation |
+| Visual recipes | 240 open recipes | Custom/private recipes and scored evaluation |
 | LLM prompt brain | Interface + local server | Production setup |
 | Image generation | Manual adapter | Real adapters and storage |
 | ERP integration | Overview docs | Connectors, field mapping, deployment |

README.zh-CN.md

@@ -4,7 +4,7 @@
 
 这不是一个“提示词搬运合集”。提示词库适合看案例、找灵感；Image2 Ads Studio 解决的是广告业务里更靠前的问题：客户只说一句需求时，系统如何解析意图、匹配广告模板、提取视觉配方、通过 LLM 重构 prompt，并输出可测试、可记录、可继续集成的作图方案。
 
-这是 **Community Edition**。它开源 Prompt Agent 框架、本地 Web 工作台、当前完整的 220 条业务模板、140 条视觉配方、LLM Prompt Brain 接口、确定性 prompt 校验、参考图策略和公开案例库；商业版提供 ERP 集成、真实生图链路、生产案例库和私有模板治理。
+这是 **Community Edition**。它开源 Prompt Agent 框架、本地 Web 工作台、当前完整的 220 条业务模板、240 条视觉配方、LLM Prompt Brain 接口、确定性 prompt 校验、参考图策略和公开案例库；商业版提供 ERP 集成、真实生图链路、生产案例库和私有模板治理。
 
 ![Workflow](docs/assets/github-workflow.svg)
 
@@ -51,7 +51,7 @@ Image2 Ads Studio 的目标是把前期需求整理成结构化广告作图方
 ## 功能亮点
 
 - 覆盖广告图文高频场景：门头、海报、展架、背景板、导视、形象墙、本地促销、电商主图、产品广告、商业摄影。
-- 开源模板库内置 220 条业务模板和 140 条视觉配方。
+- 开源模板库内置 220 条业务模板和 240 条视觉配方。
 - 支持 LLM Prompt Brain，对规则 prompt 做二次优化。
 - 支持用户上传图的多模态分析边界。
 - 支持确定性 prompt 校验，减少“高级感”“某某风格”这类模糊表达。
@@ -97,7 +97,7 @@ flowchart LR
 | 本地 Web UI | 包含 | 包含 |
 | 核心 Prompt Agent 框架 | 包含 | 包含 |
 | 业务模板 | 220 条开源模板 | 定制私有模板与治理 |
-| 视觉配方 | 140 条开源配方 | 定制私有配方与评分评测 |
+| 视觉配方 | 240 条开源配方 | 定制私有配方与评分评测 |
 | LLM Prompt Brain | 接口与本地服务 | 生产化部署 |
 | 图片生成 | 手动测试 Adapter | 真实生图 Adapter 与存储 |
 | ERP 集成 | 文档说明 | 字段映射、连接器、部署支持 |

docs/launch-kit.md

@@ -38,7 +38,7 @@ This first Community Edition focuses on the prompt-planning layer before image g
 - Local Web workbench
 - Core prompt agent framework
 - 220 advertising business templates
-- 140 visual recipes
+- 240 visual recipes
 - LLM prompt brain interface
 - Deterministic prompt validation
 - Reference image policy
@@ -75,7 +75,7 @@ I built this as a focused prompt agent for advertising image generation, not as
 
 The current Community Edition takes a business brief and reference image policy, parses the intent, retrieves advertising templates and visual recipes, uses an LLM prompt brain to rewrite the output, and validates that the final prompt contains concrete composition, lighting, material, typography, and source-reference instructions.
 
-It includes a local Web workbench, 220 business templates, 140 visual recipes, and 100 public gallery cases. It does not call an image generation API yet; the current workflow is to copy the optimized prompt into Image2 for manual testing.
+It includes a local Web workbench, 220 business templates, 240 visual recipes, and 100 public gallery cases. It does not call an image generation API yet; the current workflow is to copy the optimized prompt into Image2 for manual testing.
 ```
 
 ## Reddit
@@ -103,7 +103,7 @@ The Community Edition includes:
 
 - local Web UI
 - 220 advertising templates
-- 140 visual recipes
+- 240 visual recipes
 - LLM prompt brain interface
 - 100 image-to-prompt gallery cases
 - manual Image2 workflow
@@ -136,7 +136,7 @@ The useful part is not a single "magic prompt". The system separates:
 - reference image policy
 - exportable prompt records
 
-The current open-source version includes 220 templates, 140 visual recipes, and 100 public gallery cases. The gallery uses public upstream examples as attributed structure references, but the prompts are regenerated and debranded through the agent flow.
+The current open-source version includes 220 templates, 240 visual recipes, and 100 public gallery cases. The gallery uses public upstream examples as attributed structure references, but the prompts are regenerated and debranded through the agent flow.
 
 Repo: https://github.com/kwistzzqq-byte/image2-ads-studio
 ```
@@ -166,7 +166,7 @@ Repo: https://github.com/kwistzzqq-byte/image2-ads-studio
 
 - 本地 Web 工作台
 - 220 条广告业务模板
-- 140 条视觉配方
+- 240 条视觉配方
 - 100 个一图一 prompt 的公开 gallery 案例
 - 参考图策略
 - 手动 Image2 测试流程
@@ -196,7 +196,7 @@ Outline:
 1. 为什么不是再做一个提示词合集
 2. 广告作图和泛创意 prompt 的差异
 3. 核心流程：Brief -> Parser -> Template Retriever -> Visual Recipe -> LLM Brain -> Prompt
-4. 当前社区版：220 模板、140 配方、100 gallery 案例
+4. 当前社区版：220 模板、240 配方、100 gallery 案例
 5. 上游 prompt 库如何作为结构参考，而不是直接复制
 6. 为什么需要确定性校验：构图、材质、灯光、文字层级、参考图策略
 7. 后续路线：真实 Image2 Adapter、ERP 集成、案例评分、模板治理