claude-code-presets

Execute the check workflow from this AI development team preset. Use when the user writes /check, asks for check, or wants the corresponding team process in Codex.

Execute the dev workflow from this AI development team preset. Use when the user writes /dev, asks for dev, or wants the corresponding team process in Codex.

Execute the fix workflow from this AI development team preset. Use when the user writes /fix, asks for fix, or wants the corresponding team process in Codex.

Execute the plan workflow from this AI development team preset. Use when the user writes /plan, asks for plan, or wants the corresponding team process in Codex.

Execute the project-preset workflow from this AI development team preset. Use when the user writes /project-preset, asks for project-preset, or wants the corresponding team process in Codex.

Execute the review-all workflow from this AI development team preset. Use when the user writes /review-all, asks for review-all, or wants the corresponding team process in Codex.

Execute the ship workflow from this AI development team preset. Use when the user writes /ship, asks for ship, or wants the corresponding team process in Codex.

Execute the standup workflow from this AI development team preset. Use when the user writes /standup, asks for standup, or wants the corresponding team process in Codex.

系统化调试方法（语言无关）——复现、缩小范围、定位根因、修复、验证。具体工具按技术栈。

测试方法（语言无关）——TDD、测试金字塔、Mock 原则、覆盖率哲学、反模式。具体测试工具（Vitest / pytest）见所用预设。

Evaluate, adapt, or create Claude Code/Codex skills, agents, commands, hooks, and preset contributions from external repositories or internal ideas. Use when the user asks to borrow GitHub skills, compare mature AI coding team projects, import community agents, design a new skill, audit SKILL.md quality, or decide whether a third-party skill should be adopted.

Evaluate, adapt, or create Claude Code/Codex skills, agents, commands, hooks, and preset contributions from external repositories or internal ideas. Use when the user asks to borrow GitHub skills, compare mature AI coding team projects, import community agents, design a new skill, audit SKILL.md quality, or decide whether a third-party skill should be adopted.

Diagnose and improve Expo/React Native mobile performance including startup, lists, images, animations, JS thread work, bundle size, network retries, and production-build profiling.

Build a device verification matrix and validate React Native or Expo changes across Expo Go, development builds, simulators, emulators, real devices, deep links, permissions, and release builds.

Plan and execute React Native or Expo SDK upgrades, dependency migrations, prebuild/bare compatibility checks, native module updates, and post-upgrade validation.

Design and implement Expo Router navigation, auth flows, tabs, stacks, route params, and deep links for mobile apps.

Prepare Expo React Native apps for EAS builds, internal distribution, App Store submission, Google Play submission, and release gate checks.

Plan and implement React Native tests with unit tests, React Native Testing Library, Maestro flows, and native capability mocks.

Build and review React Native mobile UI with safe areas, touch states, keyboard handling, accessibility, and platform-aware polish.

Add React Native or Expo native capabilities such as camera, media library, location, notifications, files, sharing, and permissions.

Design mobile offline behavior, local persistence, cache policy, sync queues, weak-network states, and conflict handling.

架构设计方法（语言无关）——分层、模块边界、依赖方向、设计模式、何时不要抽象。具体技术栈选型见所用预设的 PRESET.md/rules。

单文件代码审查方法，6 个维度逐项检查（正确性、安全性、性能、可维护性、测试、无障碍）。方法语言无关，无障碍仅前端适用。

工作分解方法——把需求拆成模块/任务、定优先级、评风险、划 MVP。供 /plan（产品级拆模块）和 Architect-Planner（功能级拆任务）使用。

Execute the taste workflow from this AI development team preset. Use when the user writes /taste, asks for taste, or wants the corresponding team process in Codex.

UI 原型产出技能，按项目设计方向生成静态 HTML 原型，作为 Builder 的设计契约

性能优化方法（语言无关）——先测量再优化、定位瓶颈、优化热点、验证。具体技术（前端 Core Web Vitals、后端查询、AI 延迟）见所用预设。

UI 原型产出技能，按项目设计方向生成静态 HTML 原型，作为 Builder 的设计契约

单文件代码审查方法，6 个维度逐项检查（正确性、安全性、性能、可维护性、测试、无障碍）。方法语言无关，无障碍仅前端适用。

测试方法（语言无关）——TDD、测试金字塔、Mock 原则、覆盖率哲学、反模式。具体测试工具（Vitest / pytest）见所用预设。

架构设计方法（语言无关）——分层、模块边界、依赖方向、设计模式、何时不要抽象。具体技术栈选型见所用预设的 PRESET.md/rules。

工作分解方法——把需求拆成模块/任务、定优先级、评风险、划 MVP。供 /plan（产品级拆模块）和 Architect-Planner（功能级拆任务）使用。

构建 AI Agent——Claude 工具调用、智能体循环、多智能体协调（Orchestrator-Worker）、人机回路、安全边界

文档摄入管道——多格式解析（PDF/MD/HTML）、递归分块、去重、批量摄入 CLI，完整可运行代码

嵌入模型选型与优化——API vs 本地模型、批量嵌入、Redis 缓存、中文优化

LLM 系统评估框架——构建评估数据集、检索/答案质量指标、LLM-as-Judge、CI 集成自动评估

Prompt 设计与优化——RAG 模板、Few-shot、Chain-of-Thought、结构化输出、Prompt Injection 防护

RAG 管道设计与实现——数据摄入、分块、嵌入、混合检索、重排序到答案生成的完整流程，含常见问题诊断

从 LLM 可靠获取结构化 JSON——工具调用强制结构化、Pydantic 验证、失败重试、批量提取

pgvector 向量数据库——IVFFlat vs HNSW 索引、混合检索 RRF、Python asyncpg 封装、性能调优