跳到正文
cccskills
免费在 GitHub 公开

photo-to-cute-3d-toy

Create cute 3D toy assets from photo-reference manifests. Use when Codex needs to turn scene photos into 2D toy concept prompts, choose a photo-to-3D route, produce or plan GLB assets for Three.js, or maintain a repeatable personal-gift/toy-character generation workflow.

中文简介正在准备中。当前显示原文说明。

实测评分 未实测

查看安装方法

包含的文件(7)

  • SKILL.md2.6 KB
  • agents/openai.yaml260 B
  • references/delivery.md1.1 KB
  • references/manifest.md2.2 KB
  • references/prompting.md1.9 KB
  • references/routes.md3.5 KB
  • scripts/build_prompt_pack.py4.2 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Photo To Cute 3D Toy

Purpose

Turn a scene manifest into a repeatable pipeline:

  1. map photos to roles such as face, outfit, pose, mood, and mainPhoto;
  2. write a 2D cute-toy concept prompt pack;
  3. choose a 3D generation route;
  4. deliver frontend-ready concept images and GLB assets without leaking privacy, IP, or unsupported completion claims.

Workflow

  1. Read the project scene manifest first. If none exists, create one from references/manifest.md.
  2. Inspect the referenced photos and state the role of each photo before generating anything.
  3. Write a prompt pack using references/prompting.md.
  4. Choose a route using references/routes.md.
  5. Deliver assets using references/delivery.md.
  6. Validate outputs. If a concept image or GLB was not actually generated, say so plainly and provide the next command/service step.

Required Boundaries

  • Use original style language such as soft plush, blind-box proportions, designer collectible toy, or cute stylized figurine; do not name or imitate protected toy IPs.
  • Do not use celebrity likenesses, artist names, brand logos, readable protected marks, or copyrighted stage text in generated assets.
  • Treat private photos as sensitive. If an external service is used, report the exact service/model and which input images were uploaded.
  • Do not download large model weights, install CUDA stacks, or run long local jobs without explicit user approval.
  • Do not claim a GLB is complete unless a real file exists and passes at least a basic GLB/GLTF validation.
  • Do not modify the frontend unless the user explicitly assigns frontend integration. Hand off paths and manifest fields instead.

Scripts

Use scripts/build_prompt_pack.py to validate one scene and emit a Markdown prompt pack:

python scripts/build_prompt_pack.py `
  --manifest .\toy-scenes.manifest.json `
  --project-root . `
  --scene sence5 `
  --out .\sence5-prompt-pack.md

The script only uses the manifest and filesystem checks. It does not upload photos or call AI services.

References

  • Manifest schema and example: references/manifest.md
  • 2D concept prompt templates: references/prompting.md
  • 3D route selection: references/routes.md
  • Frontend delivery and validation: references/delivery.md

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Route CEO-first multi-window work into owner-led role loops. Use for 总控/CEO、架构/CTO、内容主编、角色派发与复用、模型路由、来源窗口回调、role-windows 台账、skill 命中统计或 multi-window loop engineering。

中文简介正在准备中。当前显示原文说明。

Dirtytrii/codex-skills112026年10月9日 更新

Read-only first workflow for diagnosing application incidents by collecting service, log, network, configuration, and functional evidence before proposing any fix.

中文简介正在准备中。当前显示原文说明。

Dirtytrii/codex-skills112026年10月9日 更新

Run user-authorized black-box web security tests and evidence-based reports. Use for 黑盒安全测试、渗透测试、数据泄露、未授权访问、公开 JS/API、登录爆破防护、CORS 或安全响应头检查。

中文简介正在准备中。当前显示原文说明。

Dirtytrii/codex-skills112026年10月9日 更新

Choose the narrowest reliable browser surface across Codex Browser, signed-in Chrome, Playwright CLI, and platform scripts. Use for browser automation, logged-in workflows, UI inspection, screenshots, form filling, publish previews, or frontend QA.

中文简介正在准备中。当前显示原文说明。

Dirtytrii/codex-skills112026年10月9日 更新

提议并执行 rubric 或 bucket 升级。两种模式:**完整 rubric bump**(最高风险动作,5 步强制 + 跨模型审核)和 **--bucket-only 轻量重校**(只换 bucket 边界,不动 rubric 公式)。**Phase 2 强制走 cheat-score-blind sub-agent 给校准池重打分**——不接受 self-scored fallback。触发词:"升级 rubric"/"bump rubric"/"更新公式"/"我想加一个维度"/"调整权重"/"重校桶"/"recalibrate bucket"。

中文简介正在准备中。当前显示原文说明。

Dirtytrii/codex-skills112026年10月9日 更新

初始化 cheat-on-content 内容校准工作区。仅在用户明确开始该工作流,或其子 skill 确认缺少 .cheat-state.json 时使用;不响应一般项目初始化、安装诊断或 Codex 首次会话。

中文简介正在准备中。当前显示原文说明。

Dirtytrii/codex-skills112026年10月9日 更新

查看 Dirtytrii 的全部技能

报告此技能的问题