Claude Skills by superduke

Guidelines for building AI-powered features into applications. Covers tool-use agent patterns, Anthropic API integration, prompt engineering for app-internal agents, and best practices. Referenced by planner and generators. Not directly user-invocable.

Full ganvil build pipeline. Takes a brief product description and autonomously builds the complete application using the Planner-Generator-Evaluator architecture with Sprint decomposition. Handles frontend, backend, and fullstack projects. Use when you want to build a complete application from a simple description.

后台管理系统还原设计标准（restore 模式专用）。以"与效果图的还原度 Fidelity" 为唯一验收口径，替代 frontend-design 的"原创性 Originality"。提供还原契约 结构、三层裁定模型、Fidelity 6 维评分标准与校准示例、电力企业后台设计锚点。 供 restore-planner 与 frontend-evaluator（restore 分支）使用。

Evaluate existing code using the ganvil evaluation framework. Automatically determines if the project is frontend, backend, or fullstack and applies the appropriate evaluator with four-dimension scoring. Use to QA-check any existing project or codebase.

Frontend design standards and aesthetic guidelines. Provides grading criteria for visual design quality, originality, craft, and functionality. Referenced by the planner and frontend evaluator. Not directly user-invocable.

Generate a product specification from a brief description. Classifies the project as frontend/backend/fullstack, defines features, creates a visual design language (for frontend), and decomposes into sprints with testable acceptance criteria. Use when you only need a product spec without building the application.

后台管理系统效果图还原模式（admin 专用）。以效果图 + spec 为输入，高保真还原 生成后台管理系统 demo。两阶段（restore:plan 产还原契约 + restore:build 按契约 生成）+ 双闸门（G1 契约审查 / G2 吻合度报告），以还原度 Fidelity 为唯一验收 口径。区别于 /ganvil:build 的"一句话自由生成"。