Claude Skills by mingfer

Activate for creative brainstorming, exploratory analysis, and idea exploration. Trigger keywords: 头脑风暴, 探索, 创意, 调研, 初步分析, 调研报告, 构思, 可能性分析, 创意发散, 讨论, 技术细节, 分析. NOT for: confirmed requirements (use m-req), confirmed design (use m-design), task planning (use m-plan). Output: exploration report to docs/explorations/.

代码审查技能，在提交代码前自动触发。专注于代码质量、安全性和可维护性。 Trigger keywords: 代码审查, code review, 审查代码.

Activates when transitioning from vetted requirements (m-req output) to solution design. Trigger signals: 设计, 概要设计, 详细设计, 技术方案, 架构, 怎么做, 实现方案, system design, architecture design, ADR. Produces Markdown design documents under docs/designs/, indexed via index.md.

Use when users need to build professional documents with AsciiDoc + Mermaid + Pandoc + Word templates. Signals: 技术文档, 方案文档, 测试报告, 用户手册, AsciiDoc, Word 导出, Mermaid 图表, 文档构建, 写文档. Outputs to project docs/ directory, generates Word .docx via Asciidoctor → Pandoc pipeline.

Activates when an execution plan (m-plan output) is confirmed and ready to execute. Trigger keywords: 开始执行, 执行任务, start execution, run plan, 执行计划. Uses TaskCreate to manage task state and dependencies. Reads docs/plans/features/<feature>/index.md and wave-N.md, dispatches sub-agents, runs review loops, and tracks all state via TaskUpdate.

Activate for simple bugs, field additions, minor API tweaks, UI copy/layout adjustments. Trigger keywords: bug, 修复bug, 修复, 加字段, 加接口字段, 加个字段, 简单修改, 小改动, 快速修复, 改下, fix, quick fix, 改一个, 改文案, 调布局, 改样式, 改UI, 界面调整. NOT for: new features (use m-chat), complex refactoring (use m-chat), unclear requirements (use m-chat).

Activates when starting a new project or analyzing an existing project. Trigger keywords: 初始化项目, 初始化, init, 分析项目, 项目诊断, 扫描项目. Initializes directory structure, copies CLAUDE.md.template, or analyzes existing project state.

Activate when requirements (m-req) and design (m-design) are both confirmed, and you need to generate an executable task plan for AI sub-agents. Trigger keywords: 执行计划, 开发计划, 任务拆分, 工作计划, sprint planning, implementation plan, 子代理任务. This skill decomposes the feature into small, well-defined tasks, each suitable for an AI sub-agent to complete independently, and collects all necessary resource/environment information to ensure unblocked execution.

Use when users describe a new feature, analyze business requirements, write user stories, define acceptance criteria, discuss product needs, author a PRD, plan a process optimization, or say things like "我想做个...功能". Signals include vague requirements, technical solution proposals, missing acceptance criteria, undefined user roles, or unclear business rules.

Activates when m-design is confirmed and a feature testing plan is needed. Trigger keywords: 测试计划, 测试用例, test plan, test cases, 回归测试, regression test. Reads design doc (testing strategy) and requirement doc, outputs Test Plan and Test Cases to docs/tests/features/<feature>/.