国内 MCP 服务推荐（飞书 / 高德 / 12306）

Read before running tests. Detailed instructions for single, standard unit tests (fast), full suites (slow), handling authentication, and obtaining the API Token. Tests must be run when a job is complete.

Docker containers, images, volumes, networks (CLI + Dockerode)

Use whenever adding, writing, drafting, or fixing a release note in the Actual Budget repo. This is the changelog entry that ships with a code change, living as a Markdown file in `upcoming-release-notes/`. Trigger for asks like "add a release note", "write the changelog entry", "add the release note for this PR/change", "create the upcoming release note", or any time finishing a user-facing change in this repo where a release note is the natural next step. These notes are published for humans to read, so they must be short, plain-language, and free of technical detail. Writing them like a commit message or with implementation jargon produces entries that get flagged and rewritten in review.

一种追问式会话，用来从用户那里挖掘片段，也就是各类写作素材（主张、小场景、尖锐句子、半成形想法），并追加到单一文档中，作为未来文章的原始素材。适用于用户想在施加结构前发展想法，或提到 “fragments”、“ideate”、“raw material” 写作素材时。

以自选路径风格，把文章塑造成一段节拍旅程。用户从原始素材中选择起始节拍，你只写这一段，然后提供下一步转向选项，逐个节拍推进，直到文章自然结束。适用于用户已有原始素材，并想把它组装成叙事而不是论证时。

通过对话会话，把一份原始素材 Markdown 文件塑造成文章：起草候选开头，逐段扩展，并在每一步讨论格式（列表、表格、提示块、引用）。适用于用户有一堆笔记、片段或粗稿，并希望把它变成可发布作品时。

ClickHouse database patterns, query optimization, analytics, and data engineering best practices for high-performance analytical workloads.

ClickHouse database patterns, query optimization, analytics, and data engineering best practices for high-performance analytical workloads.

Find refactoring targets in Python code using pyscn - duplicate code (clones), overly complex functions, and dead code. Use when user asks about refactoring, code duplication, complexity hotspots, unreachable code, or cleaning up a codebase.

Kubernetes cluster, pod, node, and workload monitoring. Use when analyzing K8s health, resource optimization, pod failures, OOMKills, scheduling, or security posture. Also use for Kubernetes operational events like pod restarts, OOM events, evictions, and cluster event history. Trigger: "Kubernetes pods", "K8s cluster health", "OOMKill", "pod restarts", "container CPU", "namespace resource usage", "over-provisioned pods", "privileged containers", "pod placement", "K8s node capacity", "running containers by cluster", "workload scheduling", "pod evictions", "K8s labels and annotations", "kubernetes events", "pod restart events", "OOM events", "K8s event history". Do NOT use for explaining existing queries, product documentation questions, AWS-specific resource queries, service-level RED metrics, distributed tracing, or log analysis — use the relevant skill instead.

Migrates React projects and components from Radix UI to Base UI. Use when asked to migrate from radix, move to base-ui, convert radix primitives, or switch a shadcn project's base library. Handles single components ("migrate accordion") and whole projects.

Use when querying or extending GitNexus's PDG control/data-dependence surface (the `pdg_query` MCP tool, CDG/REACHING_DEF edges), or reasoning about "what controls X" / "where does Y flow" / guard clauses. Examples: "what guards this statement?", "trace this variable within the function", "why is the pdg_query result empty?", "add a CDG query".

Use when the user asks about GitNexus itself — available tools, how to query the knowledge graph, MCP resources, graph schema, or workflow reference. Examples: "What GitNexus tools are available?", "How do I use GitNexus?"

Generate Salesforce Flows using the MCP tool execute_metadata_action. Use when the user asks to create, build, or generate a flow — including Screen, Autolaunched, Record-Triggered (before/after-save), Scheduled. Also trigger for flow-like requests such as "when a record is created", "trigger daily at", "send an email when", "update the field when", "automate", "workflow", or "flow XML/metadata". This is the only skill for Salesforce Flow generation.

Unified Azure OpenAI model deployment skill with intelligent intent-based routing. Handles quick preset deployments, fully customized deployments (version/SKU/capacity/RAI policy), and capacity discovery across regions and projects. USE FOR: deploy model, deploy gpt, create deployment, model deployment, deploy openai model, set up model, provision model, find capacity, check model availability, where can I deploy, best region for model, capacity analysis. DO NOT USE FOR: listing existing deployments (use foundry_models_deployments_list MCP tool), deleting deployments, agent creation (use agent/create), project creation (use project/create).

Orchestrates macro-level UI architecture, complex Jetpack Compose structures, and screen-to-screen flows. Use this skill to implement adaptive layouts (WindowSizeClass), migrate to type-safe Navigation Compose graphs, design cross-module Design Systems, wire complex screen-level state (MVI/MVVM), or orchestrate deep links and multi-screen workflows.

Browser automation — setup the bsession environment, fetch info from a website (one-shot), create scripted automations (one-shot or recurring), or debug existing sessions. Works from any repo.

Investigates distributed application performance using PostHog APM (OpenTelemetry span) data via MCP. Use when the user asks about service traces, slow HTTP/database spans, error spans, error-rate trends or spikes, latency distributions, trace IDs, or span attributes — not AI observability traces or product logs. Uses posthog:query-apm-spans, posthog:apm-trace-get, posthog:apm-spans-sparkline, posthog:apm-services-list, posthog:apm-attributes-list, and posthog:apm-attribute-values-list.

Query PyTorch CI, GitHub Actions, HUD, Grafana, and infrastructure metrics. Use when users ask about CI duration, job failures, queue times, workflow trends, runner health, dashboard data, or PyTorch infrastructure metrics.

Set up and manage local Kubernetes clusters using KIND (Kubernetes IN Docker). Use when testing Kubernetes applications locally or developing cloud-native workloads.

Kubernetes deployment, cluster architecture, security, and operations. Use for manifests, Helm charts, RBAC, network policies, operators/CRDs, PodSecurityStandards, troubleshooting, and production best practices.

Serve a quantized or unquantized LLM checkpoint as an OpenAI-compatible API endpoint using vLLM, SGLang, or TRT-LLM. Use when user says "deploy model", "serve model", "start vLLM server", "launch SGLang", "TRT-LLM deploy", "AutoDeploy", "benchmark throughput", "serve checkpoint", or needs an inference endpoint from a HuggingFace or ModelOpt-quantized checkpoint. Do NOT use for quantizing models (use ptq) or evaluating accuracy (use evaluation).

Core browser usage guide. Read this before running any browser commands. Covers the snapshot-and-ref workflow, navigating pages, interacting with elements (click, fill, type, select), extracting text and data, taking screenshots, managing tabs, handling forms and auth, waiting for content, running multiple browser sessions in parallel, and troubleshooting common failures. Use when the user asks to interact with a website, fill a form, click something, extract data, take a screenshot, log into a site, test a web app, or automate any browser task.

Simplifies code for clarity. Use when refactoring code for clarity without changing behavior. Use when code works but is harder to read, maintain, or extend than it should be. Use when reviewing code that has accumulated unnecessary complexity.

Jest testing patterns, factory functions, mocking strategies, and TDD workflow. Use when writing unit tests, creating test factories, or following TDD red-green-refactor cycle.

Core component library and design system patterns. Use when building UI, using design tokens, or working with the component library.

When the user wants to optimize their GitHub profile, README, or project discoverability. Trigger phrases include "GitHub README," "README optimization," "GitHub profile," "GitHub stars," "GitHub discoverability," "awesome lists," or "GitHub marketing."

Structures git workflow practices. Use when making any code change. Use when committing, branching, resolving conflicts, or when you need to organize work across multiple parallel streams.

A portable, framework-agnostic SEO system for any React or React Native-for-web frontend. Centralizes site metadata in one constants module, derives canonical URLs from a single base, builds per-route metadata (title, description, canonical, Open Graph, Twitter/X cards), generates...

A portable, framework-agnostic discipline for the write path of any React or React Native app using a query/cache layer. Codifies the optimistic-update lifecycle (cancel in-flight queries → snapshot every affected cache → patch instantly → roll back verbatim on error → invalidate on...

A portable, framework-agnostic field-side observability system for any React or React Native app. Establishes one typed event taxonomy (canonical event-name constants, never inline strings), a best-effort non-blocking provider fan-out so a failing or absent analytics provider can never...

A portable, framework-agnostic discipline for type safety at the network edge of any React or React Native app. Establishes one typed API client as the single fetch boundary, a parse-don't-validate rule that turns wire JSON into trusted domain types before it enters the app, a single...

A portable, framework-agnostic architecture style for any React or React Native frontend. Organizes apps into feature modules with page/screen directories, a strict server-state vs UI-state split, barrel-only cross-module imports, co-located styles, and clear component-promotion rules....

Direction and intent for frontend design. Use with design when defining purpose, tone, domain, color world, and review bar; includes cross-domain lens from cinema, architecture, marketing, UX, automotive, industrial design.

Visual philosophy and art-direction for frontend. Use when creating high-concept work, campaigns, or when the user asks for a visual philosophy, manifesto, or unmistakable art-like aesthetic.

Wenn es um Leasing mobiler Arbeitsgeräte: Laptops, Tablets, Smartphones in Leasingrecht Praxis geht: prüft Frist, Form, Zuständigkeit, Rechtsweg und Sofortmaßnahmen; liefert eine Fristen- und Risikoampel mit Sofortschritten.

Testing practices for this monorepo — choosing between unit and end-to-end tests, where test files live, how tsconfig.test.json fits in, how to drive Playwright, to to write/adapt UI/end-to-end tests, and which test runner (vitest vs jest) to use. MUST be invoked BEFORE any code search or file reads whenever the user asks to add, write, or extend a test, add coverage, create a test file, set up testing in a new package, mentions vitest/jest, or refers to an existing `*.test.ts` file. Applies equally to unit tests and to end-to-end / UI tests — the skill routes to a Playwright-MCP-first discovery flow for the latter, which differs from normal code-search workflow.

让 AI 助手具备设计能力，支持本地 ComfyUI、MeiGen 云端、OpenAI 兼容 API 三种后端，覆盖从灵感搜索到成品输出的完整创意工作流。

Review a GitHub PR's UI/UX changes by launching the MLflow web app, driving a headless agent-browser over the changed surfaces, and writing a Markdown UI-review comment body (findings + screenshots) for the workflow to post.

Audit and fix visual hierarchy, spacing, color, and depth in web UIs. Use when the user mentions "my UI looks off" (or amateur/unprofessional), "fix the design", "Tailwind styling", "color palette", "visual hierarchy", "design system", "spacing scale", or "component styling". Also trigger when building consistent design tokens, creating dark mode themes, improving data-visualization clarity, or polishing UI details before launch. Covers grayscale-first workflow, constrained design scales, shadows, and component styling. For typeface selection, see web-typography. For usability audits, see ux-heuristics.

Use when creating or editing documentation pages in docs/content/guides/ - covers YAML frontmatter, page structure, framework-specific example embedding, writing style, and sidebar registration

Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes — Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/PRD asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".

Testing standards for LoongCollector: unit tests, e2e tests, benchmarks. Reference when writing or reviewing tests.

Use when working in apps/mobileAppYC. Covers React Native architecture, navigation, Redux Toolkit state management, and mobile-specific conventions.

Use when extracting, formatting, refreshing, or investigating a read-only formatted snapshot of the installed Cursor.app bundle under .cursor-app-formatted; includes git-ignore rules, snapshot generation workflow, and the rule to inspect formatted code without patching either the snapshot code or the installed app.

Use this skill when writing new features, fixing bugs, or refactoring code. Enforces test-driven development with 80%+ coverage including unit, integration, and E2E tests.