Development

前端 UI/UX 设计开发专家代理，擅长将设计转化为精美的界面实现。当用户需要以下帮助时使用：(1) 创建新的 UI 组件 (2) 改进界面视觉效果 (3) 实现动画和交互 (4) 响应式设计 (5) 设计系统构建 (6) 即使没有设计稿也能创造美观界面。触发词包括：「UI」「界面」「组件」「样式」「动画」「好看」「美化」「设计」等前端视觉相关请求。

快速代码库搜索代理，专注于在当前项目中定位代码、文件和模式。当用户需要以下帮助时使用：(1) 查找特定代码位置 (2) 搜索文件和模块 (3) 理解代码结构 (4) 追踪函数调用链 (5) 查找相似实现。触发词包括：「在哪里」「哪个文件」「找一下」「有没有」「搜索」「定位」等代码搜索类问题。

Generate hand-drawn style diagrams (architecture, flowcharts, system design) as .excalidraw.json files. Use when user wants diagrams, mentions Excalidraw, or needs Mermaid-to-visual conversion.

Install and configure the Workflow Development Kit for resumable, durable AI agent workflows with step-level persistence, stream resumption, and agent orchestration.

Add privacy-focused web analytics with Vercel Web Analytics. Track page views, visitors, and custom events with zero configuration.

Create and run durable workflows with steps, streaming, and agent execution. Covers starting, resuming, and persisting workflow results.

--> --- name: bio-mpn-research-assistant description: Myeloproliferative neoplasm research support across JAK2, CALR, MPL, PPM1D, megakaryocyte biology, fibrosis, clonal evolution, and translational oncology. Use when synthesizing MPN literature, structuring hypotheses, interpreting MPN datasets, summarizing biomarkers, or connecting molecular findings to clinical and experimental follow-up. tool_type: mixed primary_tool: Unknown measurable_outcome: Execute skill workflow successfully with va...

--> --- name: bio-computational-software-development description: Full-stack computational software development for biomedical and life science applications. Use when building or refactoring research software, LLM-enabled analysis platforms, scientific web apps, data pipelines, RAG systems, evaluation harnesses, or package-quality Python services for translational research. tool_type: mixed primary_tool: Unknown measurable_outcome: Execute skill workflow successfully with valid output within ...

Capture exceptions, add context, create performance spans, and use structured logging with Sentry.

Manage React state in URL query parameters with nuqs. Covers Suspense boundaries, parsers, clearing state, and deep-linkable dialogs.

Use structured logging with Pino throughout your application. Covers log levels, context, and workflow-safe logging patterns.

Use Better Auth for client and server-side authentication. Covers session access, protected routes, sign in/out, and fetching user data.

Track custom events and conversions with Vercel Web Analytics. Covers common events, form tracking, and development testing.

Create a structured format for documenting feature requirements as user stories. JSON files with testable acceptance criteria that AI agents can verify and track.

Discover and follow recipes via MCP resources for setup guides, skills, and cookbooks. The meta-skill for using fullstackrecipes effectively.

Sync React state to URL query parameters for shareable filters, search, and deep-linkable dialogs with nuqs.

Complete subscription system with Stripe, Vercel Flags for plan configuration, webhook handling for syncing subscription state to Postgres, usage tracking, and billing portal integration.

Workflow from differential expression results to functional enrichment analysis. Covers GO, KEGG, Reactome enrichment with clusterProfiler and visualization. Use when taking DE results to pathway enrichment.

Complete subscription billing system with Stripe integration, feature flags for plan gating, webhook handling, and billing portal.

Syntax highlight code blocks with Shiki. Supports server-side rendering in RSC and automatic light/dark theme switching.

Add Shadcn UI components with dark mode support using next-themes. Includes theme provider and CSS variables configuration.

Configure Resend for transactional emails like password resets and email verification.

Discover novel miRNAs and quantify known miRNAs using miRDeep2 de novo prediction from small RNA-seq data. Use when identifying new miRNAs or performing comprehensive miRNA profiling with discovery.

Set up automated agent-driven development with Ralph. Run AI agents in a loop to implement features from user stories, verify acceptance criteria, and log progress for the next agent.