Skills
96,770 packaged capabilities and playbooks for agents. Install any of them with one command.
Cloud & DevOps
Start, query, and stop a network-specific TAO inference microservice ({network_arch}-inference-microservice) by delegating container execution to the appropriate platform skill. Handles container image resolution, job-payload JSON construction, and the service registry. Use when the user wants to run inference on a TAO model checkpoint using a microservice container, deploy a TAO inference endpoint, or stop a running inference container.
Host setup for TAO GPU backends. Checks and, after user approval, installs NVIDIA driver branch 580, CUDA Toolkit 13.0, and NVIDIA Container Toolkit 1.19.0 for Docker/local-Docker and Kubernetes GPU worker hosts. The `--check-only` path works on any Linux distribution; `--install` automates debian-family (Ubuntu/Debian/Pop!_OS/Mint/Zorin/Raspbian), rhel-family (Fedora/RHEL/Rocky/AlmaLinux), and suse-family (openSUSE/SLES) hosts, and prints actionable manual-install steps for everything else.
This skill should be used when the user asks to "write a Tiltfile", "configure Tilt", "set up live update", "debug Tilt", "add a resource to Tilt", "optimize Tilt builds", "view Tilt logs", "restart a Tilt resource", or mentions Tiltfile, tilt up, tilt ci, tilt down, live_update, docker_build, custom_build, k8s_resource, local_resource, or Kubernetes local development with Tilt.
Build, review, debug, configure, migrate, teach, or plan Turborepo 2.11 monorepos with current docs and a full usage guide. Use for turbo, create-turbo, turbo.json, turbo.jsonc, tasks dependsOn ^build, inputs outputs, deferred input hashing (jit, dependencyOutputs), task tags and --filter=tag:, envMode strict, remote cache TURBO_TOKEN TURBO_TEAM, local cache eviction (cacheMaxAge, cacheMaxSize), filters --affected, futureFlags, turbo watch, turbo query, turbo prune --production and Docker, package configurations, boundaries, generators, devEngines.packageManager, bun/pnpm/nub/aube workspaces, experimental Rust/Python/Go (Cargo, uv, go.work) toolchains, agentGuidance AGENTS.md block, Next.js Vite Vitest Playwright oxlint oxfmt task graphs, and CI.
Managing platform versions and Renovate annotations in the homelab. Use when: (1) Adding a new version entry to versions.env, (2) Configuring Renovate to track a new dependency, (3) Debugging why Renovate ignores or mis-detects a version, (4) Understanding annotation syntax for versions.env, (5) Adding container image tracking to YAML files, (6) Configuring package rules or grouping in Renovate. Triggers: "versions.env", "renovate annotation", "renovate not updating", "add version", "renovate ignore", "datasource", "extractVersion", "package rule", "automerge", "renovate validate", "dependency tracking", "version management".
Use this skill when the user wants to deploy, run, debug, tear down, or call the REST API of the RTVI-CV 2D detection / tracking microservice. Trigger when the user says things like 'deploy rtvi-cv', 'start warehouse 2d', 'add a stream', 'check rtvi-cv health', or 'stop the perception container'. Not for VLM, embedding, or analytics — use the matching vss-* skill
Deploy and operate the RTVI-CV-3D microservice as MV3DT (`MODE=mv3dt`): per-camera DeepStream perception plus BEV Fusion over calibrated cameras. Supports the bundled sample dataset, custom video files, and RTSP streams, and chains to `vss-generate-video-calibration` when calibration is missing. Use `vss-deploy-profile` for the full warehouse blueprint and `vss-deploy-detection-tracking-2d` for single-camera 2D detection.
Use this skill when deploying, operating, or integrating the VSS 3.2 GA RT-Embed Video Embedding microservice. Covers Docker Compose bring-up, GPU and storage prerequisites, the `/v1` REST API (file uploads, text and video embeddings, live RTSP streams, health and metrics), Redis/Kafka/OTel integration, common failure modes, and teardown
Integration skill for Alibaba Cloud Yunxiao DevOps. Provides full DevOps capabilities, including code management, pipelines, project management, organization management, artifact repositories, application delivery, and test management. Use when users need to interact with Yunxiao, including: (1) code repository operations (branches, files, merge requests); (2) creating and running pipelines; (3) managing projects and work items; (4) managing organization members; (5) querying artifact repositories; (6) deploying and releasing applications; (7) managing test cases. Triggers: Yunxiao, Codeup, Flow, Projex, AppStack, DevOps.
Diagnose Alibaba Cloud Cloud Firewall VPC firewall provisioning failures, route policy configuration failures, and closure pre-check risks using read-only CloudFirewall, CBN, VPC, STS, and ActionTrail APIs. Use this skill when troubleshooting VPC firewall creation failures, route policy/drainage configuration failures, firewall status stuck in configuring, or assessing route and ACL impact before closing VPC firewall drainage.
Alibaba Cloud LoongCollector / SLS installation, collection onboarding, Pipeline config management and validation, machine groups, permission troubleshooting, and Lens queries. HARD RULE: for matching requests, the first tool MUST load this skill before any SSH probe, directory setup, checklist/file write, or cloud read. Triggers: "安装 LoongCollector", "ECS 安装采集器", "自建 Linux 主机安装 LoongCollector", "ACK 安装 loongcollector", "自建 K8s 部署采集", "从安装到能查到日志", "SLS 日志采集接入", "SLS 日志采集接入相关的事", "修改采集配置", "改采集配置", "采集配置校验", "validate_pipeline.py", "SLS 机器组", "新建 Logtail Pipeline 采集配置", "Logtail Pipeline", "ClusterAliyunPipelineConfig", "SLS Lens 查询", "无数据排查", "心跳异常", "SLS 采集权限排查", "SLS 权限排查", "阿里云 CLI 凭证没有 SLS 操作权限", "Logtail", "iLogtail", "AgentSight", "Agentloop", "input_agentsight", "eBPF Runtime", "ebpf-event".
This skill should be used when users need to manage AWS Support cases via CLI. It handles listing cases (recent 2 weeks, unresolved), viewing case details with bilingual display (English-Chinese line by line), creating new cases with auto-detected service/category, replying to cases, and attachment handling. Triggers on requests mentioning "AWS support", "support case", "工单", "support ticket", or AWS technical support inquiries.
Cloud service provider privacy assessment framework. Covers ISO 27018 cloud privacy controls, CSA STAR certification, SOC 2 Type II evaluation, shared responsibility model mapping, data residency verification, and cloud-specific privacy risk analysis.
Cloudflare Vectorize, the globally distributed vector database for semantic search, RAG, and AI applications on Workers. Covers index creation with fixed dimensions and distance metric (cosine, euclidean, dot-product), insert/upsert/query/delete/list vector operations, metadata and namespace filtering, topK search, document chunking and batch ingestion pipelines, and embedding generation via Workers AI (@cf/baai/bge-base-en-v1.5) or OpenAI (text- embedding-3-small/large). Use when creating a Vectorize index, wiring a vectorize binding, inserting embeddings, building a RAG or semantic search pipeline, running wrangler vectorize commands, or debugging metadata index timing errors, dimension mismatches, filter syntax problems, or insert vs upsert confusion.
Guidance for setting up Git repositories with automatic web deployment via post-receive hooks. This skill applies when configuring bare Git repositories, setting up web servers to serve pushed content, creating Git hooks for deployment automation, or implementing push-to-deploy workflows.
Guides enterprise data center portfolio planning and execution—multi-site capacity roadmaps, investment prioritization (build, expand, refresh, exit, colo vs owned), portfolio RAID and dependency management across DC programs, stage-gate governance, capex/opex alignment, regional and resiliency strategy, and steering-committee reporting. Use when prioritizing several DC initiatives, harmonizing site plans over 3–5 years, tracking a portfolio of hall builds and refreshes, or aligning facilities/IT/finance on DC investments—not for a single hall MEP design (data-center-design-execution-lead), host-level utilization (data-center-compute-supply-efficiency), generic software programs (technical-program-manager), or cloud IaC (infrastructure-engineer). For executing approved MW/rack delivery on schedule, use senior-data-center-capacity-delivery-manager.
Design and audit data residency and regional operating models covering customer region commitments, data classification, storage, processing, backups, logs, analytics, search, AI providers, subprocessors, support/admin access, replication, failover, deletion, migrations, contracts, and proof. Use when customers or regulators require clarity about where data lives and moves.
MongoDB and PostgreSQL database administration. Databases: MongoDB (document store, aggregation, Atlas), PostgreSQL (relational, SQL, psql). Capabilities: schema design, query optimization, indexing, migrations, replication, sharding, backup/restore, user management, performance analysis. Actions: design, query, optimize, migrate, backup, restore, index, shard databases. Keywords: MongoDB, PostgreSQL, SQL, NoSQL, BSON, aggregation pipeline, Atlas, psql, pgAdmin, schema design, index, query optimization, EXPLAIN, replication, sharding, backup, restore, migration, ORM, Prisma, Mongoose, connection pooling, transactions, ACID. Use when: designing database schemas, writing complex queries, optimizing query performance, creating indexes, performing migrations, setting up replication, implementing backup strategies, managing database permissions, troubleshooting slow queries.
Deploy a Decentraland scene to a World (personal 3D space using a DCL NAME or ENS domain). Covers worldConfiguration setup, Places listing opt-out, and common deployment errors. Use when the user wants to deploy to a World, publish to a personal space, or use a DCL NAME/ENS domain. Do NOT use for Genesis City LAND deployment (see deploy-scene).
Docker best practices for VPS deployments in 2026. Use this skill when writing Dockerfiles, docker-compose.yml, or managing containers on a Linux VPS. Covers multi-stage builds, uv-based Python images, Xvfb/virtual display containers, volume management, security hardening, and service orchestration with Docker Compose v2. NEVER use legacy docker-compose v1 (hyphenated), old Python base images, or root containers.
Pre-deploy security checklist for Postgres with Electric. Checks REPLICATION role, SELECT grants, CREATE on database, table ownership, REPLICA IDENTITY FULL on all synced tables, publication management (auto vs manual with ELECTRIC_MANUAL_TABLE_PUBLISHING), connection pooler exclusion for DATABASE_URL (use direct connection), and ELECTRIC_POOLED_DATABASE_URL for pooled queries. Load before deploying Electric to production or when diagnosing Postgres permission errors.
Runs the Pacu AWS exploitation framework end-to-end — session and credential setup, IAM enumeration, automated privilege-escalation scanning via iam__privesc_scan, and persistence/backdooring modules like iam__backdoor_users_keys. Use for authorized AWS penetration tests or red-team engagements, assessing blast radius from a compromised IAM credential, or generating attacker telemetry for purple-team detection testing.
Set up automated CI failure detection and fixing using Claude Code. Use when you want to create a GitHub Actions workflow that automatically analyzes workflow failures, applies fixes for common issues, and opens issues for complex problems.
Use when interacting with GitLab via the glab CLI: creating/reviewing merge requests, managing issues, monitoring CI/CD pipelines, making API calls, or performing any GitLab operation from the terminal. Triggers on glab, gitlab cli, merge request, MR create, pipeline status, ci lint.
Cloudflare Tunnel (cloudflared add-on) HTTPS for Home Assistant OS (HAOS): give a HAOS instance a real HTTPS URL. Use when the user wants HTTPS or remote access for Home Assistant, or mentions cloudflared or Cloudflare Tunnel. Also consult it before suggesting DuckDNS, port forwarding, or HA-native SSL for a HAOS box.
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.