claudeskills.page

450+ skills · Free · MIT licensed

Expert skills for Claude Code

474 installable SKILL.md playbooks, from Postgres and React to firmware, game engines and product planning. Each gives Claude a senior practitioner's workflow, real commands, known pitfalls and a way to verify the result, and loads only when the task calls for it.

Browse all 474 skills Start with the essentials

Popular: Systematic Debugging, Postgres Indexing, Accessibility Audit, OWASP Top 10 Review, PDF Toolkit

curl -fsSL https://claudeskills.page/install.sh | sh -s -- rate-limiting
Example: Claude Code loading the rate-limiting skill only when the request needs it.
Skills
474
Categories
25
Lines of instructions
125k
With scripts or references
207

Directory

Browse by discipline

25 categories in 5 areas, from Claude Code itself to Postgres, game engines, firmware and PowerPoint decks.

Search the full directory →

Agent & workflow

58 skills

How Claude plans, codes, reviews and automates

Build software

129 skills

Stacks, languages and platforms

Data & AI

80 skills

Models, pipelines, storage and evidence

Ship & operate

124 skills

Quality, infrastructure and security

Beyond code

83 skills

Product, design, documents and growth

Everything at once

2.5 MB

All 474 skills in one zip, organized by name and ready to unpack into a skills folder.

curl -fsSLO https://claudeskills.page/downloads/all-skills.zip
Download all-skills.zip →
By technology

New in October 2026

5 new disciplines, 56 new skills

Game engines, microcontrollers, product work, research and everyday automation, written to the same standard as the rest of the catalog.

The best skills by use case →

Editors' selection

Skills worth installing first

The ones that change how Claude behaves most visibly, chosen from every category.

Browse curated collections →
★ No. 020

Spec-Driven Development

Requirements, design, tasks, code — each traced to the last

Claude researches in plan mode, then writes requirements, design and a task list as files in the repo and stops for approval at each gate. Implementation goes one task at a time with tests named after requirement IDs, ticked tasks carry commit hashes, deviations update the design before coding continues, and a final trace table proves every requirement has a passing test.

spec-driven-development 199 lines
★ No. 093

Temporal Workflows

Durable processes that survive deploys and replays

Claude keeps Temporal workflow code deterministic and pushes every side effect into idempotent activities with explicit timeouts, heartbeats and non-retryable business errors. It models sagas with compensations, uses signals, queries and updates correctly, continues long loops as new, and treats every workflow code change as a versioning decision proven by replaying stored production histories in CI.

temporal-workflows 261 lines · 2 files
★ No. 157

Prometheus Alerting

Pages that mean act now, routed to one thread

Claude sorts signals into page, ticket or dashboard before writing a single rule, then builds recording rules, symptom-based alerts with for durations, minimum-traffic guards, severity labels and runbook links. Rules are unit tested with promtool and checked by a bundled linter, and Alertmanager gets grouping and inhibition so one outage sends one notification.

prometheus-alerting 249 lines · 2 files
★ No. 220

Statistical Testing

The right test, an effect size, honest corrections

Claude picks the hypothesis test from a decision table keyed on outcome type, number of groups and pairing, checks the assumptions that actually matter such as independence of units, and reports effect sizes with confidence intervals ahead of p-values. A bundled script applies Bonferroni, Holm, Benjamini-Hochberg and BY corrections side by side, and results are written as plain-language statements with limits.

statistical-testing 174 lines · 2 files
★ No. 288

Rendering Strategies

Static where you can, per-request only where you must

Claude inventories every route first: data sources, how stale each may be, whether it is per-user, traffic share and page count. It then picks the most static mode the facts allow, maps it to the framework's real config, keeps personalized bits out of shared caches, and proves the plan with build output, cache-hit headers and a cross-user leak test.

rendering-strategies 207 lines
★ No. 335

Landing Page Design

Pages with a point of view, not a SaaS template

Claude commits to an art direction rooted in your product before writing CSS, orders sections by the questions a skeptical visitor asks, and builds a hero with one message, one action and a real product visual as the LCP element. A bundled audit catches the tells that make pages look machine-made: gradient blobs, emoji feature grids, centered everything and fake proof.

landing-page-design 161 lines · 2 files
★ No. 396

Stripe Payments

Webhook-driven checkout that never double-charges

Claude builds Stripe flows the way payment engineers do: prices resolved on the server, one Customer per account, idempotency keys tied to the business operation, and fulfillment driven only by signature-verified webhooks that tolerate retries and out-of-order delivery. It then proves the flow with Stripe CLI triggers, decline and 3DS cards, and test clocks.

stripe-payments 253 lines · 2 files
★ No. 434

Firmware OTA Updates

Updates that survive power cuts and bad builds

Claude designs OTA so a device always boots something that works: A/B slots, bootloader signature checks, anti-downgrade rules, and a self-test that must pass before the image is confirmed. It writes the confirm and rollback code for ESP-IDF and MCUboot, generates release manifests, runs a power-cut test matrix, and plans staged rollouts with numeric halt gates.

firmware-ota 212 lines · 2 files
★ No. 472

Scheduled Jobs

Cron jobs safe to rerun, overlap-proof and watched

Claude designs the job to be rerun safely before choosing a schedule, then wires it into a systemd timer, cron, launchd or a Kubernetes CronJob with catch-up, jitter and a time limit. A bundled wrapper adds a portable single-instance lock, UTC logs and healthcheck pings, and every job ships with a heartbeat monitor that alerts when it silently stops running.

scheduled-jobs 212 lines · 2 files

How skills work

Install dozens. Pay for one at a time.

Skills use progressive disclosure. Claude Code reads only each skill's short description at startup. The full playbook enters context when your request matches it, and bundled scripts and references load only if a step needs them.

What are Claude Code skills? →
  1. Frontmatter~100 tokens · always loaded
    name: postgres-indexing
description: Designs and audits PostgreSQL indexes
  from real query plans… Use when a query is slow
  or EXPLAIN shows a Seq Scan on a large table…
  2. SKILL.md body~2–5k tokens · on match

    Workflow, decision tables, commands, pitfalls, and how to verify the result.

  3. references/ and scripts/0 tokens until used

    Deep catalogs and deterministic helpers. Scripts run; only their output enters context.

Editorial standard

What every skill includes

Each skill was written to the same specification before it was published.

Read the editorial standards →

  • A selective trigger

    The description says what the skill does and exactly when to load it, so it fires on the right requests and stays out of the way otherwise.

  • A real workflow

    Numbered steps in the order a senior practitioner works, including what to inspect before changing anything.

  • Decisions with numbers

    Tables and thresholds — LCP under 2.5 s at p75, lock timeouts, contrast ratios — instead of “it depends”.

  • Known pitfalls

    The mistakes that only show up in production, called out before Claude makes them.

  • A verification step

    Every skill ends with how to prove the work is right, and what to do when the check fails.

  • Fully portable

    Only standard frontmatter, so the same zip works in Claude Code, on Claude.ai and through the Skills API.

Install

One command, no restart

The installer downloads the zip and unpacks it into your skills folder. Claude Code watches that folder, so the skill is available in your next message.

~/.claude/skills/
Personal — every project on this machine
.claude/skills/
Project — commit it and your team gets it
Every install method, explained → 
# install one or more skills for your user
curl -fsSL https://claudeskills.page/install.sh | sh -s -- \
  systematic-debugging code-review
Verify with /skills inside Claude Code

Collections

Curated stacks for one kind of work

21 hand-ordered bundles. Each installs with a single command or downloads as one zip.

All 21 collections →
More stacks

Guides

Install, write and debug skills

All guides →

FAQ

Questions, answered

The official skills documentation covers every field and setting.

What is a Claude Code skill?

A skill is a folder with a SKILL.md file: YAML frontmatter (a name and a description of when to use it) followed by instructions. Claude Code keeps only the short description in context and loads the full instructions when a request matches, so you can install many skills without slowing the agent down.

How do I install a skill from this site?

Run curl -fsSL https://claudeskills.page/install.sh | sh -s -- skill-name to install into ~/.claude/skills/, or add --project to install into the current repository's .claude/skills/ folder. You can also download the zip from any skill page and unzip it into either folder. Claude Code picks up new skills without a restart.

Are these skills free? What license are they under?

Yes. Every skill is free to download and use, including commercially, under the MIT license. There is no account, registry client or paid tier.

Which Claude Code skills should I install first?

Start with the workflow skills that change every session: systematic-debugging, task-decomposition-plans, code-review and commit-craft. The Starter Pack bundles ten of them in one command, and the best skills by use case lists the picks for frontend, backends, security, DevOps, documents and AI work.

How many skills are in the directory?

474 skills across 25 categories, covering software engineering, infrastructure, data and AI, and work beyond code such as product management, research, documents and writing. The full list is available as plain text and JSON, and every skill can be downloaded together as one zip.

Is this an official Anthropic website?

No. Claude Skills is an independent, community-run directory. It is not affiliated with or endorsed by Anthropic. For the official specification, see the Claude Code skills documentation.

Do these skills work on Claude.ai and the Claude API too?

Yes. Every skill uses only the portable Agent Skills frontmatter fields (name, description, license), so the same zip can be uploaded to Claude.ai or the Skills API. Skills that rely on running scripts need an environment with code execution enabled.

How is this different from other skill directories?

Every skill here was written to one quality bar: a concrete workflow, decision tables, real commands, known pitfalls and a verification step, typically 150–350 lines rather than a paragraph of generic advice. Each page shows the complete SKILL.md and every bundled file before you download anything.

The full catalog is free and open

Every skill is MIT licensed. No account, no registry client.

Browse 474 skills
esc