国内 MCP 服务推荐（飞书 / 高德 / 12306）

Operate 20+ penetration-testing tools (Nmap, Nuclei, SQLMap, FFUF, Hashcat, and more) through structured workflows with consistent output handling.

Go from reverse engineering to a working exploit: stack/heap/kernel pwn workflows with pwntools, libc-database, ROP, and stabilization from CTF to authorized remote targets.

Search what was said in existing screenpipe, Codex, Claude, Cursor, and Gemini CLI chats, then continue or steer one exact chat when the user explicitly asks. Use for requests about finding another agent conversation, recalling what an agent said, or sending work to one.

Automate browser interactions, test web pages and work with Playwright tests.

Authoritative reference for how docs in this repo (and 5 other Netdata-org repos) become published pages on `learn.netdata.cloud`. Covers the `<repo>/docs/.map/map.yaml` source-of-truth (the actual lever -- filesystem path is irrelevant for routing), the live `ingest/ingest.py` orchestrator in the learn repo (NOT the legacy `ingest.js`), frontmatter injection, slug rules, sidebar autogeneration, MDX escape rules, versioning, the 4-mechanism redirect stack, the 6 source repositories, the every-3-hours CI ingest, Netlify deploy, and the `part_of_learn=True` opt-in for files hand-authored in the learn repo. Use when adding/moving/renaming/deleting a docs page; when a page on Learn looks wrong; when wondering whether to edit a doc here or in the learn repo; when reading `ingest.py`, `sidebars.js`, `docusaurus.config.js`, `static.toml`, `LegacyLearnCorrelateLinksWithGHURLs.json`, `netlify.toml`, the `<!--startmeta` blocks in `.mdx` files, or the workflows `ingest.yml` and `daily-learn-link-check.yml`.

Configures alerting policies in Terraform for Google Kubernetes Engine (GKE) clusters, workloads, and services using PromQL and Google Cloud Managed Service for Prometheus. Use when writing, analyzing, validating, or deploying Terraform alerting policies to monitor GKE service latency, traffic, error rates using Multi-Window Multi-Burn-Rate SLO alerts, memory saturation, and cluster health such as CrashLoopBackOff and Node NotReady conditions. Don't use for non-GKE compute runtimes such as standalone Compute Engine VMs or standalone Cloud Run services without GKE.

Design and build AI agents for any domain. Use when users: (1) ask to "create an agent", "build an assistant", or "design an AI system" (2) want to understand agent architecture, agentic patterns, or autonomous AI (3) need help with capabilities, subagents, planning, or skill mechanisms (4) ask about Claude Code, Cursor, or similar agent internals (5) want to build agents for business, research, creative, or operational tasks Keywords: agent, assistant, autonomous, workflow, tool use, multi-step, orchestration

Use when writing, reviewing, restoring, trimming, or auditing prose in the deepseek-harness repo, including deciding where documentation or comments are required across Markdown, JSDoc, code and test comments, prompts, descriptions, diagnostics, and CLI or UI strings.

MUST USE for frontend/web UI/UX/visual work: building, styling, redesigning pages/components, React setup, performance audits, visual QA, taste, and polish. Routes four rulesets: design taste router and brand references; perfection for Playwright/Chromium Lighthouse/Core Web Vitals; ui-ux-db palettes/fonts/guidelines; designpowers personas/accessibility/critique/handoff; plus curl-only lazyweb real-app-screen research and the beui.dev interaction catalog. Triggers: frontend, UI, UX, design, redesign, styling, layout, animation, motion, interaction, micro-interaction, make it feel alive, premium, luxury, minimal, brutalist, Awwwards, DESIGN.md, mockup, React, Lighthouse, accessibility, WCAG, Core Web Vitals, looks generic, make it pretty, like X brand, lazyweb, design research.

Design or review infrastructure, deployment, CI/CD, Docker, Kubernetes, health checks, rollback, feature flags, production readiness, and mobile release workflows. Use for DevOps, platform, container, pipeline, or release tasks.

Multi-source deep research using firecrawl and exa MCPs. Searches the web, synthesizes findings, and delivers cited reports with source attribution. Use when the user wants thorough research on any topic with evidence and citations.

Use this skill when retrieving Jira tickets, analyzing requirements, updating ticket status, adding comments, or transitioning issues. Provides Jira API patterns via MCP or direct REST calls.

Use when working on the Chat Customizations editor — the management UI for agents, skills, instructions, hooks, prompts, MCP servers, and plugins.

Uses Chrome DevTools via MCP for efficient debugging, troubleshooting and browser automation. Use when debugging web pages, automating browser interactions, analyzing performance, or inspecting network requests. This skill does not apply to `--slim` mode (MCP configuration).

Launches, monitors, debugs, and collects one run-scoped AIPerf Kubernetes benchmark without changing its workload semantics. Use after perf.yaml and aiperf-config.yaml have been configured for a smoke-tested Dynamo deployment.

Record upgrade instructions alongside a Prisma Next breaking-change PR, so downstream consumers (users of `@internal/*` and authors of Prisma Next extensions) can apply the matching code translation automatically via the published upgrade skills. Use when you have refactored framework code and the test suite went red in `examples/` or `packages/3-extensions/`, when you fixed those red tests by editing the substrate, when you are told to "record upgrade instructions for this PR", or when you made a breaking change to Prisma Next that downstream consumers will need help migrating across.

Guide for migrating viewport media queries (@media, useMedia, and screen:-prefixed responsive props) to container queries in Sentry's frontend. Use when migrating responsive layout to container queries, replacing @media/useMedia/screen: breakpoints, refactoring styled responsive components to Container/Flex/Grid primitives, or working on the DE container-query migration.

Generates optimized prompts for AI tools. Activates only when the user explicitly asks to write, fix, improve, or adapt a prompt for a specific AI tool (LLM, Cursor, Midjourney, image AI, video AI, coding agents, etc.). Does not activate for general conversation, coding tasks, document writing, or other non-prompt-engineering work.

Operate Famulor assistants, communication history, campaigns, knowledge, automations, telephony, and workspace administration through its hosted MCP server.

Scrape web pages using Scrapling with anti-bot bypass (like Cloudflare Turnstile), stealth headless browsing, spiders framework, adaptive scraping, and JavaScript rendering. Use when asked to scrape, crawl, or extract data from websites; web_fetch fails; the site has anti-bot protections; write Python code to scrape/crawl; or write spiders.

Implements supply chain integrity verification for container builds with the in-toto framework: generating signing keys, defining a supply chain layout, recording pipeline steps as signed link metadata, verifying before deployment, enforcing at Kubernetes admission, and integrating with SLSA. Use when attesting CI/CD pipeline steps, proving an image followed the approved build process, or enforcing provenance at admission. Keywords: in-toto, layout, link metadata, step, inspection, SLSA, provenance, admission. Do not use for signing and verifying images with Cosign - use implementing-image-provenance-verification-with-cosign.

Audits container and pod configuration for escape-enabling misconfiguration using the Kubernetes Python client - privileged flags, dangerous capability grants, host path mounts, shared namespaces, and CVE-2022-0492 style cgroup abuse. Use when sweeping a cluster for workloads that could break out, producing a posture report, or checking configuration before enforcement is switched on. Keywords: privileged, hostPath, hostPID, capabilities, CVE-2022-0492, cgroup, kubernetes python client, posture audit. Do not use for runtime syscall-based detection - use detecting-container-escape-attempts. '

Implements eBPF-based runtime observability and in-kernel enforcement in Kubernetes with Cilium Tetragon, monitoring process execution, file access, network connections, and syscalls, and blocking dangerous calls at the kernel level. Use when deploying Tetragon to detect or block syscalls such as ptrace, mount, and unshare, enforcing kernel-level policy, or adding low-overhead runtime detection to a cluster. Keywords: Tetragon, Cilium, eBPF, TracingPolicy, kprobe, enforcement, process lineage. Do not use for Falco-based detection - use detecting-container-runtime-threats-with-falco.

Uses Calico's own policy CRDs beyond the upstream Kubernetes API - GlobalNetworkPolicy, HostEndpoint, NetworkSet, policy tiers, and DNS-based egress rules - applied and audited with calicoctl. Use when a policy must span namespaces or protect the host itself, when egress has to be expressed by domain name, or when ordering policies into tiers. Keywords: calicoctl, GlobalNetworkPolicy, HostEndpoint, NetworkSet, tier, DNS egress, order. Do not use for portable upstream NetworkPolicy - use implementing-network-policies-for-kubernetes; for installing Calico and writing standard policy with it use implementing-kubernetes-network-policy-with-calico.

Deploys OPA Gatekeeper via Helm as a Kubernetes admission controller and writes ConstraintTemplates with Rego plus instantiated Constraints to validate, mutate, or deny resource requests at admission time. Use when enforcing custom policy-as-code at admission on Kubernetes v1.24+, blocking non-compliant workloads before scheduling, or expressing a rule that built-in controls cannot. Keywords: Gatekeeper, ConstraintTemplate, Constraint, Rego, admission webhook, audit, mutation. Do not use for the standard pod security profiles that Pod Security Admission already covers - use implementing-pod-security-admission-controller.

Configures and operates the Kubernetes Pod Security Admission (PSA) controller that enforces Pod Security Standards: namespace enforce/audit/warn labels, cluster-wide defaults via AdmissionConfiguration, exemptions for usernames, runtime classes and namespaces, version pinning, and troubleshooting pods the controller rejected. Use when wiring PSA up on a cluster, setting cluster-wide default enforcement, exempting system namespaces, debugging why a pod was rejected or why enforcement is not firing, or reading PSA audit and warning output. Keywords: Pod Security Admission, PSA, admission controller, AdmissionConfiguration, pod-security.kubernetes.io labels, enforce audit warn, exemptions, kube-apiserver. Do not use for choosing which security profile a workload needs - use implementing-kubernetes-pod-security-standards.

Detects and prevents privilege escalation inside Kubernetes pods by combining admission control (OPA policies), runtime monitoring (Falco), and audit log analysis of security contexts, Linux capabilities, and syscall patterns. Use when investigating a pod running as root or privileged, hardening workloads against in-pod escalation, or hunting for containers exceeding their intended scope. Keywords: allowPrivilegeEscalation, runAsRoot, capabilities, securityContext, OPA, Falco, audit log. Do not use for escalation through RBAC and service-account permissions - use auditing-kubernetes-rbac-privilege-escalation.

Assesses the security posture of the etcd cluster backing Kubernetes: encryption at rest, TLS peer and client transport, access control, backup encryption, and network isolation. Use when auditing or hardening a control plane, reviewing whether Secrets are encrypted at rest, or protecting etcd backups, since etcd stores Secrets, RBAC policy, and ConfigMaps in plaintext by default. Keywords: etcd, EncryptionConfiguration, encryption at rest, peer TLS, snapshot, backup, control plane. Do not use for broad cluster-wide CIS checks - use performing-kubernetes-cis-benchmark-with-kube-bench.

Turns kube-bench output into a finished CIS Kubernetes Benchmark audit: interpreting PASS/FAIL/WARN per control, judging which failures are genuine on a managed cluster, writing remediation, and packaging evidence for SOC 2 or PCI DSS. Use when conducting a scheduled compliance audit, triaging kube-bench results, deciding which controls are not applicable on EKS, GKE, or AKS, or producing hardening evidence for an auditor. Keywords: CIS Kubernetes Benchmark, control plane, remediation, compliance evidence, SOC 2, PCI DSS, managed cluster exception. Do not use for installing and running the tool - use benchmarking-kubernetes-with-kube-bench.

Chooses and applies the correct Kubernetes Pod Security Standard (Privileged, Baseline, Restricted) for a workload: what each profile forbids, how to map existing workloads to a profile, which securityContext fields must change, and how to plan a PodSecurityPolicy-to-PSS migration without breaking running pods. Use when deciding which pod security profile a namespace or workload should run under, auditing which workloads would fail Restricted, planning a PSP migration, or mapping pod security posture to a compliance control. Keywords: Pod Security Standards, PSS, Privileged, Baseline, Restricted, securityContext, runAsNonRoot, drop ALL capabilities, seccomp RuntimeDefault, PSP migration. Do not use for configuring the admission controller that enforces these profiles - use implementing-pod-security-admission-controller.

Installs Calico as the cluster CNI and writes standard Kubernetes NetworkPolicy under it, covering default-deny baselines, policy ordering and precedence, service-account-based selectors, and verifying that policy is genuinely being enforced. Use when adopting Calico as the enforcement CNI, establishing a default-deny baseline, or debugging why a NetworkPolicy is not taking effect under Calico. Keywords: Calico CNI, NetworkPolicy, default deny, policy order, Felix, service account selector. Do not use for Calico-only CRDs such as GlobalNetworkPolicy or DNS egress - use implementing-container-network-policies-with-calico; for CNI-agnostic policy use implementing-network-policies-for-kubernetes.

Installs and runs the kube-bench tool against a Kubernetes cluster as a Job, DaemonSet, or standalone binary, selecting the correct benchmark version and targets (control plane, etcd, kubelet, worker nodes) and emitting JSON or JUnit output for pipelines. Use when setting kube-bench up for the first time, choosing which benchmark version and node targets to run, wiring it into CI, or troubleshooting skipped or misdetected checks. Keywords: kube-bench, DaemonSet, --benchmark, --targets, JSON output, JUnit, CI integration. Do not use for interpreting the findings or producing an audit report - use performing-kubernetes-cis-benchmark-with-kube-bench.

Exploits privileged pods, host mounts, runC CVEs, and exposed Docker sockets to break out of a container and reach the underlying host during an authorized container-security assessment. Use when executing an approved breakout test, demonstrating the real impact of a privileged or hostPath workload, or validating that escape mitigations actually hold. Keywords: container breakout, privileged, hostPath, docker.sock, runC CVE-2019-5736, CVE-2024-21626, release_agent, nsenter. Do not use for defensive detection of these techniques - use detecting-container-escape-attempts.

Hardens the Docker daemon (dockerd) through /etc/docker/daemon.json with user namespace remapping, TLS client authentication, seccomp profiles, and CIS Docker Benchmark controls such as icc, no-new-privileges, and live-restore. Use when securing a Docker host's daemon to prevent privilege escalation, breakout, or lateral movement, or when auditing daemon settings against CIS requirements. Keywords: dockerd, daemon.json, userns-remap, no-new-privileges, icc, live-restore, TLS socket. Do not use for hardening images and per-container runtime flags - use hardening-docker-containers-for-production.

Hardens Dockerfiles, images, and per-container runtime settings against the CIS Docker Benchmark v1.8.0: non-root users, dropped capabilities, read-only root filesystem, seccomp and AppArmor profiles, and minimal multi-stage builds, validated with docker-bench-security, Hadolint, and Dockle. Use when preparing a container or Dockerfile for production, or auditing images and runtime flags against CIS Docker controls. Keywords: Dockerfile, USER, --cap-drop, read-only rootfs, seccomp, AppArmor, multi-stage, Hadolint, Dockle. Do not use for the Docker daemon's own configuration - use hardening-docker-daemon-configuration.

Evaluates Kubernetes cluster security by actively simulating attacker techniques against the API server, kubelet, etcd, pods, RBAC, network policy, and secrets, using kube-hunter, Kubescape, peirates, and manual kubectl exploitation to find paths to cluster compromise. Use for an authorized penetration test or hands-on validation that controls actually stop an attacker. Keywords: kube-hunter, Kubescape, peirates, kubelet 10250, anonymous auth, token theft, lateral movement, cluster takeover. Do not use for a configuration-only compliance audit - use performing-kubernetes-cis-benchmark-with-kube-bench.

Writes portable upstream Kubernetes NetworkPolicy YAML - default-deny-all, DNS egress, namespace and pod selector rules - that works on any conformant CNI such as Calico or Cilium. Use when segmentation must stay CNI-portable, introducing a default-deny posture, or restricting east-west traffic between pods and namespaces without depending on a vendor CRD. Keywords: NetworkPolicy, default deny, podSelector, namespaceSelector, ingress, egress, CNI portable. Do not use for Calico-specific resources - use implementing-kubernetes-network-policy-with-calico.

Runs Docker Bench for Security, the open-source CIS Docker Benchmark audit script, across host configuration, daemon settings, images, and runtime configuration, then interprets pass/fail/warn output and remediates the common failures. Use when auditing Docker hosts for CIS compliance, scheduling recurring container assessments, or validating runtime hardening controls after a change. Keywords: docker-bench-security, CIS Docker Benchmark, audit script, pass fail warn, host configuration, remediation. Do not use for applying the daemon hardening itself - use hardening-docker-daemon-configuration.

Comprehensive guide for building with Prisma 8 (Prisma Next), the contract-first data layer. Use whenever working on Prisma code in a project that uses it — authoring or editing the data contract (contract.prisma, PSL, TypeScript builders), migrations, queries (db.orm / db.sql), runtime wiring (db.ts, middleware, DATABASE_URL), build-tool integration, Supabase / RLS, reading PN-* structured errors, or filing feedback — and for orientation questions like "what is Prisma Next" or comparisons to other ORMs. Signals that this skill applies: @internal/* imports, prisma.config.ts, contract.prisma / contract.json / contract.d.ts, the prisma-next CLI, PN-* error codes. Does not apply to Prisma ORM 7 or earlier (schema.prisma + @prisma/client projects).

Official CCXT MCP (Model Context Protocol) server — connect an AI agent to 100+ cryptocurrency exchanges and prediction markets for market data, balances, and opt-in trading. Covers installing the server in Claude Desktop/Code, Cursor, VS Code and other MCP hosts; configuring named accounts and API keys (kept local, never seen by the model); the capability tiers and safety rails; sandbox/testnet trading; and how to drive the tools. Use when the user wants an agent to fetch prices/order books/candles, check balances/orders/positions, place or cancel orders, or explore prediction markets across exchanges.

Cherry Studio first-party tool and bundled-shell routing for general agents. For straightforward local work in shell-capable sessions, run JS/TS with `bun <file>` and one-off JS tools with `bun x`; run Python with `uv run [--with <pkg>] python` and one-off Python CLIs with `uvx`; search with `rg`. Load this guide before changing project dependencies, deciding whether a tool should be ephemeral or reusable, reading or converting local Office/PDF files, coordinating or delegating across Agent Sessions, or using Cherry-owned web/browser, knowledge, persistent memory, schedules/notifications, IM channels, image generation, artifact reporting, managed CLI, skill, or MCP-server-registration capabilities—even if the user names no tool. Consult it before shell/file workarounds; live tool schemas are authoritative.

How to write, run, and debug compiler tests and fourslash (LSP) tests in the TypeScript repository. Covers test file formats, directives, markers, baseline management, and the fourslash verification API.

Validates all code changes before committing by running format, lint, type, and unit test checks. Use after making backend (Python) or frontend (TypeScript) changes, before committing or finishing a work session.

Use when the user wants to manage or troubleshoot tooling, remotes, sync, build, deploy, logs, CI/CD, or operational workflows for an existing Twenty app.

Guide for exposing PostHog product endpoints as MCP tools. Use when creating new or updating API endpoints, adding MCP tool definitions, scaffolding YAML configs, or writing serializers with good descriptions. Covers the full pipeline from Django serializer to generated TypeScript tool handler.

Manage stacked branches and pull requests with the gh-stack GitHub CLI extension. Use when the user wants to create, push, rebase, sync, navigate, or view stacks of dependent PRs. Triggers on tasks involving stacked diffs, dependent pull requests, branch chains, or incremental code review workflows.