Skills
96,770 packaged capabilities and playbooks for agents. Install any of them with one command.
Cloud & DevOps
Expert knowledge for the Hugging Face Hub ecosystem including HTTP/curl API usage, Python/JavaScript SDKs, Spaces apps (Gradio, Docker), model formats (GGUF, PEFT/LoRA), datasets (pandas, Polars, DuckDB), automation (webhooks, Jobs), and security features. Use when working with Hub APIs, huggingface_hub Python code, Spaces apps, inference providers, model/dataset repositories, or Hub automation.
Enable and configure Amazon Macie via AWS CLI/Terraform to discover, classify, and protect sensitive data (PII, financial data, credentials) in S3 using ML and pattern matching, including discovery jobs, custom data identifiers, allow lists, and EventBridge-based remediation. Use when setting up S3 data classification, cloud DLP, or auditing S3 for unprotected sensitive data.
Run comprehensive joelclaw health checks, including postboot/daily checks for NAS, MinIO, Convex, Postgres, Typesense, Inngest, Panda/Flagg authority split, k8s, worker, Redis, OTEL, tests, repo sync, memory pipeline, and disk. Use when: 'system health', 'postboot check', 'daily system check', 'health check', 'is everything working', 'system status', 'how's the system', 'check everything', or at session start to orient.
Langfuse was decommissioned on 2026-07-09 — do not add Langfuse tracing or query Langfuse for usage/cost. Use this skill only to learn where usage observability lives now. Token usage and cost questions route to `joelclaw usage` and ClickHouse otel_events.
Configure LaunchDarkly Factory settings (GitHub App auto-flagging and auto-releasing) via the hosted MCP, or diagnose why a pull request was not classified / auto-flagged. Use when the user wants to turn on auto-flagging, map a GitHub repo to a LaunchDarkly project, change Factory account defaults, unmap a repo, or asks why Factory did not classify their PR
Use when driving mc, the MinIO AIStor command-line client for S3-compatible object storage, from a shell — connecting to a server with an alias, listing/copying/moving/removing objects and buckets, or configuring versioning, replication, lifecycle, encryption, policies, events, cluster administration, and AIStor Tables. Covers how an agent should authenticate non-interactively, get machine-readable JSON-lines output, build ALIAS/BUCKET/PREFIX paths, discover any command's exact flags, and stay clear of the irreversible destructive operations
Run a scenario-based DevOps, SRE, Cloud, or Platform Engineering mock interview, then deliver a brutally honest analysis and a tailored prep plan. Pulls questions configured by topic (Kubernetes, AWS, Docker, Terraform, CI/CD, Linux, Monitoring/Observability, System Design, Networking, Security, mixed), difficulty (junior 0-2y / mid 2-5y / senior 5-8y / staff 8+y), and target company tier (Service like TCS/Infosys/Wipro/Capgemini, Product like Razorpay/Zerodha/Swiggy/PhonePe/Atlassian, or MANG/FAANG like Meta/Apple/Netflix/Google/Amazon/Microsoft). Conducts the interview question-by-question in interviewer voice (no hints, no mid-interview praise), then flips to coach mode for per-question analysis, pattern detection across the session, a tier-by-tier verdict, and a 7-day prep plan. Trigger AGGRESSIVELY on: /mock-interview, "mock interview", "interview prep", "interview practice", "DevOps interview", "SRE interview", "cloud interview", "Kubernetes interview", "AWS interview", "scenario question", "system design interview", "interview me", "test me", "quiz me on K8s/AWS/etc", "help me prepare for an interview", or anytime the user expresses they have an upcoming DevOps/SRE/Cloud/Platform/Infra interview, even if they don't use the word "mock" — interview readiness is the trigger, not the exact phrasing.
Monitor and debug web and Node.js applications using OpenTelemetry, console logs, network requests, and distributed tracing. Use when the user asks about distributed tracing, correlating frontend/backend requests, OpenTelemetry, Jaeger, or monitoring application behavior.
Mailgun (Sinch) platform help — developer-first transactional email API and SMTP relay with inbound routing, webhooks, and Mailgun Optimize deliverability tools. Use when Mailgun emails not delivering, DNS records (DKIM/SPF) failing verification, inbound routing not catching messages, webhooks not firing, templates rendering wrong, Mailgun Optimize showing poor inbox placement, or API calls returning errors. Do NOT use for general email deliverability strategy (use /sales-deliverability), cross-platform email marketing (use /sales-email-marketing), or email open/click tracking strategy (use /sales-email-tracking).
Deciding what a system should report about itself, independent of the SDK that carries it out — the vendor-neutral layer over logs, metrics, traces, error handling, and product events. Triggers on code that logs, throws, catches, reports an error, tracks an event, or identifies a user, and on "log level", "capture exception", "error boundary", "breadcrumb", "trace sampling", "cardinality", "event name", "identify", "reset on logout", or "cookie consent". For which package to install, which option controls collection, or source maps and build-time tokens, use the installed vendor's own capability. Covers try/catch placement, metric cardinality, and event naming through consent.
Building MCP (Model Context Protocol) servers in TypeScript on Cloudflare Workers with the official @modelcontextprotocol/sdk: registering tools, resources, and prompts, Hono HTTP routing, StreamableHTTPServerTransport setup, authentication (API keys, OAuth, Zero Trust), Cloudflare D1/KV/R2/Vectorize integration, wrangler deployment, and testing. Prevents common failures including export syntax errors, Zod schema validation failures, memory leaks from unclosed transports, CORS misconfiguration, and missing rate limiting. Use when exposing APIs, tools, or data to LLMs as a stateless serverless MCP server on the edge. Not for Python MCP servers (see FastMCP), stateful WebSocket agents (Cloudflare Agents SDK), or local stdio CLI tools.
Upgrades Helm chart dependencies (PostgreSQL, Vault) in the Chainloop project, including vendorized charts, container images, and CI/CD workflows. Use when the user mentions upgrading Helm charts, Bitnami dependencies, PostgreSQL chart, or Vault chart. CRITICAL - Major version upgrades are FORBIDDEN and must be escalated.
Use when the user wants to transact on any Volcengine commercial product — purchase/provision it, ask what it costs, or unsubscribe from it. Covers three actions: buy (purchase, subscribe, provision, open, place an order; Chinese "开通 / 购买 / 下单 / 买 / 售卖 / 订购"), price (how much does it cost, quote, billing rules, estimated spend; Chinese "多少钱 / 价格 / 报价 / 费用 / 怎么收费 / 如何计费 / 预估费用"), and unsubscribe (cancel a subscription, refund a prepaid instance; Chinese "退订 / 退款 / 取消订阅"). Trigger even when the user does not mention "sale" but expresses a clear purchase, pricing, or cancellation intent toward a Volcengine product. Only handles these three transaction actions; anything outside them is out of scope.
Author and review Cloudflare Workers code against production best practices. Use when writing a Worker, reviewing Worker code, configuring wrangler.jsonc, or checking for Workers anti-patterns — streaming, floating promises, global state, secrets, bindings, observability.
Provides comprehensive Alibaba Cloud (Aliyun) guidance including ECS, ApsaraDB, OSS, SLB, VPC, RAM, ACK (Kubernetes), Function Compute, API Gateway, CDN, and monitoring services. Covers infrastructure provisioning with Terraform/ROS, cloud architecture design, security best practices, cost optimization, and migration strategies. Produces infrastructure code, deployment scripts, architecture diagrams, and operational procedures. Use when working with Alibaba Cloud services, designing cloud architecture on Aliyun, migrating to Alibaba Cloud, setting up Chinese cloud infrastructure, implementing multi-region deployments in China, or when users mention Alibaba Cloud, Aliyun, ECS, OSS, ApsaraDB, ACK, RDS, SLB, or Chinese cloud computing.
Browserbase cloud-headless-browser SDK expertise for TypeScript/Node. Use when code imports `@browserbasehq/sdk`, `@browserbasehq/stagehand`, `playwright-core`, or `puppeteer-core` and connects to Browserbase; when code POSTs to `api.browserbase.com/v1/sessions`, `/v1/contexts`, `/v1/extensions`, or `/v1/projects`; when sending the `x-bb-api-key` header or using `BROWSERBASE_API_KEY`/`BROWSERBASE_PROJECT_ID`; when the user asks about Browserbase sessions, contexts, proxies, downloads, uploads, recordings, live debugger URL, captcha solving, advanced stealth, verified browsers, fingerprinting, BYOS S3 storage, regions (`us-west-2`/`us-east-1`/`eu-central-1`/`ap-southeast-1`), keepAlive, the `connectUrl` CDP WebSocket, `chromium.connectOverCDP`, `puppeteer.connect({browserWSEndpoint})`, the `browser.contexts()[0]` default-context pattern, Stagehand `act`/`extract`/`observe`/`agent` primitives, the Model Gateway, or deploying serverless via Browserbase Functions.
Surfaces a single, genuinely relevant AWS Activate partner offer as optional context AFTER another AWS Startup Advisor skill has finalized its recommendation, plan, or build—never influencing the technical advice. Consulted by architect-for-startups, start-building-for-startups, agent-advisor, gcp-to-aws, heroku-to-aws, and llm-to-bedrock: once their output is final, this skill checks whether a relevant Activate offer exists for what the founder is building or migrating and, if so, appends one quiet, dismissible line with a tracked redeem link. Merit-first (offers never change the recommendation), relevance-gated to what the founder actually surfaced, at most one per response and often none, muteable. Offer content is read live from the sibling knowledge-base-for-startups skill (references/offers.md). Do NOT use for account-specific lookups (credits balance, Activate membership, application status)—send those to https://aws.amazon.com/startups.
Manage Databricks workspace connections: check current workspace, switch profiles, list available workspaces, or authenticate to a new workspace. Use when the user mentions "switch workspace", "which workspace", "current profile", "databrickscfg", "connect to workspace", or "databricks auth".
Deploy a Decentraland scene to Genesis City (LAND-based). Covers pre-deployment checklist, scene.json validation, spawn points, and common deployment errors. Use when the user wants to deploy, publish, go live, or upload to parcels they own. Do NOT use for Worlds deployment (see deploy-worlds).
Use when asked to manage UniFi network - list/restart/upgrade devices, block/unblock clients, manage WiFi networks, control PoE ports, manage traffic rules, create guest vouchers, or any UniFi controller task. Works with UDM Pro/SE, Dream Machine, Cloud Key Gen2+, or self-hosted controllers.
The complete build-to-production pipeline for Ethereum dApps. Fork mode setup, IPFS deployment, Vercel config, ENS subdomain setup, and the full production checklist. Built around Scaffold-ETH 2 but applicable to any Ethereum frontend project. Use when deploying any dApp to production.
Guidance for setting up Git-based multi-branch deployment systems with SSH access, web servers, and automated deployment hooks. This skill should be used when configuring Git repositories that deploy to multiple environments (e.g., main/dev branches), setting up SSH authentication for Git, configuring web servers to serve content from different branches, or creating post-receive hooks for automated deployments.
Harden GitHub Actions and Dependabot: create or modify workflow YAML, reusable workflows, or .github/dependabot.yml; configure dependency updates, alerts, graphs, private registries, and pull-request automation; audit security, correctness, reliability, cost, or performance.
A guided end-to-end process starting from a new machine: SSH passwordless login → install Docker → install/register GitLab Runner → get a minimal CI/CD pipeline working. Semi-automated: SKILL.md provides steps and command templates, while scripts/ offers detection, configuration, and validation scripts. The user only needs to provide the machine address and GitLab token. Trigger when the user says “configure runner,” “install gitlab runner on a new machine,” “add a runner to gitlab,” or “get CICD running.”
Guidance for setting up HuggingFace model inference services with Flask APIs. This skill applies when downloading HuggingFace models, creating inference endpoints, or building ML model serving APIs. Use for tasks involving transformers library, model caching, and REST API creation for ML models.
Design and audit incident status page operations covering component taxonomy, incident declaration, public/private updates, subscriber notifications, templates, severity, SLO impact, customer segmentation, uptime history, maintenance windows, post-incident follow-up, support coordination, and trust-preserving communications. Use when reliability incidents need accurate external communication.
Operate the joelclaw event bus, gateway, observability, and agent loop infrastructure via the joelclaw CLI. Use for: sending events, checking runs, starting/monitoring/cancelling agent loops, debugging failed runs, checking health, restarting the worker, inspecting step traces, managing subscriptions, gateway operations, OTEL queries, semantic recall. Triggers: 'joelclaw', 'send an event', 'check inngest', 'start a loop', 'loop status', 'why did this fail', 'debug run', 'check worker', 'restart worker', 'runs', 'what failed', 'subscribe', 'gateway', 'otel', 'recall', or any Inngest/event-bus/agent-loop task.
Search and cite Microsoft 365 Message Center, Microsoft 365 Roadmap and Azure Updates announcements from mc.merill.net. Use when answering questions about Message Center IDs, Roadmap items, Azure updates and retirements, Microsoft 365 change announcements, rollout dates, services, or archived posts.
Integrates Neon serverless Postgres and Vercel Postgres (built on Neon) into web and edge applications: the @neondatabase/serverless and @vercel/postgres HTTP and WebSocket drivers, pooled versus non-pooled connection strings, sslmode requirements, git-like database branching with neonctl, Drizzle ORM and Prisma integration, migrations, and point-in-time restore. Use when setting up serverless Postgres for Cloudflare Workers, Vercel Edge Functions, Next.js, or Node.js serverless functions, configuring connection pooling for edge runtimes, creating database branches, migrating from D1/SQLite or traditional Postgres to serverless Postgres, or debugging errors like connection pool exhausted, TCP connections not supported in serverless, transaction timeouts, and sslmode required.
Nginx configuration, reverse proxy, SSL/TLS, load balancing, and performance tuning. Use for: nginx, reverse proxy, load balancer, proxy_pass, ssl certificate, lets encrypt, web server, location block, upstream, server block, nginx config, certbot, hsts, gzip, rate limiting.
Implement and verify joelclaw observability on every change so failures cannot stay silent. Use when adding/updating Inngest functions, gateway channels, webhook providers, APIs, workers, or any pipeline step. Enforces canonical OTEL contract, storage path, and verification gates. Triggers on: 'o11y', 'observability', 'logging', 'otel', 'instrument this', 'silent failure', 'add telemetry', 'log this function'.
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.