Skills
96,981 packaged capabilities and playbooks for agents. Install any of them with one command.
Cloud & DevOps
Deploy OpenSearch search applications to Amazon OpenSearch Service or Amazon OpenSearch Serverless. Use this skill when the user wants to provision an OpenSearch domain or serverless collection on AWS, deploy search configurations to AWS, set up Bedrock connectors, configure IAM roles for OpenSearch, migrate a local search setup to AWS, or manage Amazon OpenSearch infrastructure. Activate even if the user says AOS, AOSS, OpenSearch Service, serverless collection, Bedrock connector, SigV4, or AWS deployment without mentioning search.
Manage Backblaze B2 cloud storage. List files, audit usage, estimate cost, clean up stale data, review security posture, and manage lifecycle rules. Use when the user mentions B2, Backblaze, object storage buckets, or storage cleanup.
Production backend systems development. Stack: Node.js/TypeScript, Python, Go, Rust | NestJS, FastAPI, Django, Express | PostgreSQL, MongoDB, Redis. Capabilities: REST/GraphQL/gRPC APIs, OAuth 2.1/JWT auth, OWASP security, microservices, caching, load balancing, Docker/K8s deployment. Actions: design, build, implement, secure, optimize, deploy, test APIs and services. Keywords: API design, REST, GraphQL, gRPC, authentication, OAuth, JWT, RBAC, database, PostgreSQL, MongoDB, Redis, caching, microservices, Docker, Kubernetes, CI/CD, OWASP, security, performance, scalability, NestJS, FastAPI, Express, middleware, rate limiting. Use when: designing APIs, implementing auth/authz, optimizing queries, building microservices, securing endpoints, deploying containers, setting up CI/CD.
Use when the user has the Bmob MCP server configured (http://mcp.bmobapp.com/mcp) and wants to perform LIVE operations against their Bmob backend cloud project from the IDE. Triggers: 'list bmob tables', 'show bmob schema', 'create bmob table', 'add a row to bmob', 'update bmob record', 'delete bmob data', '生成 bmob curl', 'design bmob schema', '上传云函数', '部署云函数', '发布云函数', '同步云函数', '同步函数', '拉取云函数', '下载云函数', '调用云函数', '执行云函数', 'invoke cloud function', 'run cloud function', '一键部署网站', '静态托管', '部署静态站点', 'deploy static site', 'deploy website'. Provides 12 tools: get_project_tables (ALWAYS call first to discover schema before any write), create_table, add_single_data, update_single_data, delete_single_data, generate_code, deploy_cloud_function, invoke_cloud_function, list_cloud_functions, get_cloud_function, deploy_static_site, mcp_endpoint_mcp_post. NOT for writing client/SDK code that will ship in the user's app — for that use bmob-database-{javascript,android,ios,swift,flutter,restful}. NOT for ACL / role design (use bmob-acl-and-roles) or cloud function development (use bmob-cloud-function-development).
Use the Browser Use cloud API to spin up cloud browsers for Clawdbot and run autonomous browser tasks. Primarily, create browser sessions with profiles (persisted logins/cookies) that Clawdbot can control. Secondarily, run task subagents for fast autonomous browser automation. Docs at docs.browser-use.com and docs.cloud.browser-use.com.
Browserless cloud browser automation service. Run headless Chrome at scale for scraping, screenshots, and PDF generation. Use for cloud browser automation, scalable scraping, or headless Chrome as a service. Triggers on browserless, headless chrome, browser service, cloud scraping, screenshot api, pdf generation, chrome as a service.
Entrypoint for managing Volcengine Flink (based on Apache Flink) workloads. Use it for developing and deploying streaming and batch jobs, diagnosing real-time issues (e.g., backpressure, checkpoint failures, OOM), troubleshooting resources, and performing SRE actions (stop/start/restart/scale/config changes). It intelligently routes requests to four specialized sub-skills and leverages flink-mcp for tool execution.
Complete knowledge domain for Cloudflare Cron Triggers - scheduled execution of Workers using cron expressions for periodic tasks, maintenance jobs, and automated workflows. Use when: scheduling Workers to run periodically, adding cron triggers to Workers, configuring scheduled tasks, testing cron handlers, combining crons with Workflows, enabling Green Compute, handling multiple schedules, or encountering "scheduled handler not found", "cron expression invalid", "changes not propagating", "handler does not export", "timezone issues" errors. Keywords: cloudflare cron, cron triggers, scheduled workers, scheduled handler, periodic tasks, background jobs, scheduled tasks, cron expression, wrangler crons, scheduled event, green compute, workflow triggers, maintenance tasks, scheduled() handler, ScheduledController, UTC timezone
Configure and troubleshoot Cloudflare DNS records and Pages custom-domain bindings. Use when a user asks to add/edit/delete DNS records, map a custom domain to Cloudflare Pages/Workers, verify propagation, or fix domain validation errors. On first use, prompt for API token setup with DNS edit permissions before making changes.
Use when inspecting or changing anything in a Cloudflare account as an agent — DNS records, WAF custom rules, bot settings, redirect, transform, cache or page rules, zone settings, tunnels, Workers, Pages, R2 or analytics. Reads go through a local read-only API token that the agent never sees. Changes are proposed for the human to make in the dashboard, or applied through an approval step when the user has deliberately stored a write token.
Build and deploy the Goldsky Compose corporate-actions / dividend-distribution example under the user's own account — a durable, idempotent distributor that pays N token holders pro-rata for a tokenized corporate action (dividend, coupon, rebate, airdrop) with an on-chain audit trail. The interesting bit: Compose orchestrates a Goldsky Turbo job-mode pipeline as an ephemeral subroutine — declaring a campaign spawns a one-shot pipeline that snapshots share-token holders at a record block, waits for it, pays each holder via a gas-sponsored wallet, then deletes the pipeline. Triggers on: 'dividend distribution', 'pay dividends onchain', 'distribute dividends to shareholders', 'corporate actions distributor', 'pro-rata payout to token holders', 'airdrop pro-rata by balance', 'cap table distribution', 'set up / deploy the dividend / corporate-actions example'. Ships pointed at shared permissionless demo contracts on Base Sepolia so there's nothing to deploy. Deploys fully in-app or via the CLI; the only runtime requirement is a `GOLDSKY_PROJECT_KEY` secret, set as the last step. For a custom/novel Compose app, use /compose. For debugging a deployed app, use /compose-doctor. For manifest/CLI/API field lookups, use /compose-reference.
Manage Coolify self-hosted PaaS from the terminal. Multi-instance contexts, application lifecycle (deploy, logs, env sync), database provisioning with backups, service management, and CI/CD automation via JSON output. Use when: deploying apps via Coolify CLI, managing servers/databases/services, syncing environment variables, automating Coolify workflows, or troubleshooting "cannot unmarshal array", 404 on service env update, context authentication failures.
Adds the five CRUD REST endpoints (all, detail/{id}, save, update, delete) for an entity to up-api and redeploys. Use when the user asks for Day 1 step 7 of the CRUD drill, or asks to add CRUD API endpoints to up-api. Needs the same entity that was used in the entity-migration step.
Dockerizes the up-app Angular project (multi-stage build served by nginx with SPA fallback) and deploys it via Docker Compose alongside up-api and the db. Use when the user asks for Day 2 step 1 of the CRUD drill, or asks to dockerize/deploy up-app.
Guides compliance with country-specific data localization requirements across key jurisdictions including Russia (242-FZ), China (PIPL Art. 40, CAC measures), India (DPDP Act), Turkey, Vietnam, and Indonesia. Covers localization assessment, architecture design, and exemption procedures. Keywords: data localization, data residency, PIPL, 242-FZ, cross-border restrictions.
Databricks Expert Engineer Skill - Comprehensive guide for data engineering, machine learning infrastructure, and permission design Use when: - Running databricks CLI commands (auth, api) - Executing SQL queries via Databricks SQL Warehouse - Working with Unity Catalog permissions - Managing Lakeflow Jobs or Delta Lake
Parse error logs, stack traces, and crash reports into a structured root cause diagnosis. Use when an application is throwing exceptions, crashing, or producing unexpected errors and you need to understand why and what to fix. Produces a structured diagnosis with error classification, stack trace walkthrough, probable root cause with confidence level, affected code path, a concrete code-level fix suggestion, and ordered next debugging steps.
Covers DeepSpeed for distributed deep learning training and I/O: ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8 training, 1-bit Adam, sparse attention, and DeepNVMe (aio_handle, gds_handle, async_io and gds operators, ds_nvme_tune, ds_report) for fast transfers between NVMe storage and host or GPU tensors. Use when configuring ZeRO stages for large-model training, enabling mixed precision or 1-bit Adam, offloading parameters or optimizer state to NVMe with ZeRO-Infinity, writing or reading tensors to files with blocking or non-blocking DeepNVMe calls, or tuning NVMe I/O settings such as block size, queue depth and parallelism. Not for plain PyTorch DistributedDataParallel without DeepSpeed.
Creates devcontainers with Claude Code, language-specific tooling (Python/Node/Rust/Go), and persistent volumes. Use when adding devcontainer support to a project, setting up isolated development environments, or configuring sandboxed Claude Code workspaces.
Automatically configure and switch to domestic mirror sources to speed up package manager installs. Use this skill when the user installs packages/dependencies (e.g. npm install, pip install, yarn install, brew install, apt install, gem install, cargo install, go get), complains about slow downloads, install timeouts or network issues, or explicitly asks about domestic mirror sources.
(NS) Generate and adapt .gitlab-ci.yml for SaaS products with backend and frontend — test, build, E2E gate, deploy by branch. Use when the user asks for GitLab CI, pipeline setup, or CI/CD bootstrap for a monorepo product. Uses references/templates/gitlab-ci.template.yml as baseline. Out of scope for pure agent-runtime repos without backend+frontend layout.
Use when uploading, releasing, promoting, or managing rollouts on Google Play. Make sure to use this skill whenever the user mentions gpc releases, upload AAB, upload APK, staged rollout, promote to production, halt rollout, gpc publish, release notes, track management, internal testing, beta release, production rollout, version code, rollout percentage, or wants to ship an Android app to any Play Store track. Also trigger when someone asks about the Google Play edit lifecycle, release validation, or how to do a phased rollout — even if they don't mention GPC by name. For metadata and listings, see gpc-metadata-sync. For CI/CD integration, see gpc-ci-integration.
Create a haloy.yaml configuration file for deploying applications with haloy. Use when the user says "create haloy config", "add haloy.yaml", "configure for haloy", "set up haloy deployment", or "prepare for haloy". Supports single-target and multi-target deployments, local Docker builds, published/private images, named images with image_key, and optional self-hosted databases. Not for creating Dockerfiles (use the dockerize skill), multi-environment deployments, or docker-compose/Kubernetes setups.
Upload local HTML or Markdown artifacts to baseurl.ai and return a shareable, password-bearing preview URL. Use this skill whenever the user asks to publish, share, host, preview, send a link for, or make accessible any local .html, .htm, .md, or .markdown file, including reports or documents you just generated. Also use it when a workflow produces renderable local content and the next useful step is a URL someone else can open. Do not use it for production deployments, long-term hosting, or non-renderable source files unless you first create a local HTML or Markdown artifact.
Wraps existing PyTorch training scripts with HuggingFace Accelerate (Accelerator class, accelerate config, accelerate launch) so the same code runs on CPU, single GPU, multi-GPU, multi-node, TPU, or Apple MPS. Covers device placement, FP16/BF16/FP8 mixed precision, gradient accumulation, distributed checkpointing, and switching between DDP, DeepSpeed ZeRO, FSDP, and Megatron backends. Use when converting a single-GPU script to multi-GPU, enabling mixed precision, configuring DeepSpeed ZeRO or FSDP, or setting up gradient accumulation. Use when one script must run on different hardware. Not for callback-heavy training loops (use PyTorch Lightning) or multi-node orchestration with hyperparameter tuning (use Ray Train).
Implements and audits GCP VPC firewall rules using gcloud, covering auditing overly permissive rules, creating restrictive ingress/egress rules, hierarchical firewall policies, and monitoring rule effectiveness with VPC Flow Logs. Use when deploying GCP workloads needing network access controls, auditing firewall configs, or responding to Security Command Center findings; not for Cloud Armor or DNS-based filtering.
Reserved and on-demand GPU cloud instances for ML training and inference. Use when you need dedicated GPU instances with simple SSH access, persistent filesystems, or high-performance multi-node clusters for large-scale training.
Introduces LinuxMirrors' positioning, capabilities, supported scope, source code and authoritative references, and verifies the official website, GitHub and current source code first when answering about project status, compatibility, script differences, recent changes or sources. Use it to understand the project and choose the right workflow; for hands-on system package source work or changing sources during Docker installation, follow the corresponding dedicated process instead — do not use this skill as a substitute for a change plan.
Plans and guides LinuxMirrors system package source changes for GNU/Linux hosts, covering distribution detection, domestic/education-network/overseas/official sources, interactive and unattended parameters, EPEL and EOL scenarios, backup, verification, troubleshooting and rollback. Use when the user explicitly wants to change apt/yum/dnf/pacman/zypper/apk/portage/Nix package sources; for Docker CE and Registry Mirror, use the dedicated Docker process instead.
Manage local Airflow environment with Astro CLI. Use when the user wants to start, stop, or restart Airflow, view logs, troubleshoot containers, or fix environment issues. For project setup, see setting-up-astro-project.
End-to-end guide for adding push notifications to .NET MAUI apps. Covers Firebase Cloud Messaging (Android), Apple Push Notification Service (iOS), Azure Notification Hubs as the cross-platform broker, an ASP.NET Core backend API, and the MAUI client wiring on every platform. USE FOR: "push notification", "FCM", "APNS", "Firebase Cloud Messaging", "Azure Notification Hubs", "device registration", "remote notification", "send push notification", "notification token". DO NOT USE FOR: local/scheduled notifications (use maui-local-notifications), general permission handling (use maui-permissions), or background tasks without notifications (use maui-app-lifecycle).
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.