Introduction

One of the hottest words in programming circles recently is non- Skills None other than.

It does not refer to the technical ability of the programmer; “Skills Pack” for AI Programming Assistants .

A set of standardized workflow modules that allows AI to transform from “writing code randomly” to “delivering according to process”.

In today’s article, I have compiled the 10 most popular Skills in the programming circle in 2026.

The code can be copied and used after installation.

Hope it will be helpful to you.

More practical projects are on my technical website: susan.net.cn/project

1. First understand what Skills is

Before talking about specific Skills, let’s spend 2 minutes to clarify one thing:What exactly are skills?

In October 2025, Anthropic launched Agent Skills Open standards. Its core idea is simple:

Package the complete process of "how to do something" into a reusable Skill file.

A Skill is essentially a folder containing a SKILL.md Files - Instructions written in Markdown, plus optional scripts and reference files. put it in ~/.claude/skills/ directory, AI can be automatically loaded and used when needed.

The core differences between Skills and traditional Prompts:

  • Prompt It's a "suggestion" - the AI ​​can refer to it or ignore it
  • Skill It’s a “process” – AI must follow steps and cannot skip them

As of July 2026, there are 1400+ Claude Code Skills, the official marketplace includes more than 658.

There are more than 15,000 officially indexed warehouses alone, and the three major marketplaces total more than 490,000 entries.

But quantity does not mean quality.

Someone specifically tested 100 community skills, and the results 70% failed.

There are no more than 10 people who can really stay in daily work.

The following 10 have been tested in actual combat.

2. Top 1: Superpowers—

It is the "engineering discipline manual" of the AI ​​programming world

GitHub Star: 243,000+

Say it clearly in one sentence: Equip AI with a complete set of software engineering methodologies, turning it from an "inspired coder" into a "reliable delivery team".

Superpowers is currently the most successful collection of Skills in the entire Skills ecosystem.

It is not a function, but a complete workflow framework containing 20+ sub-modules - TDD, systematic debugging, plan writing, code review, parallel Agent distribution, Git Worktree management... Each is an independent Skill that can be triggered independently.

core concept: Process over Prompt (process is greater than prompt word).

The core sub-skills:

① brainstorming⭐ most commonly used

This is the entry Skill and the most critical one. Its core logic is:AI is never allowed to write a line of code without explicit user approval. AI, like Socrates, only asks one question at a time, gradually clarifies the needs, and proposes 2-3 solutions for you to choose from.

② subagent-driven-development (sub-agent-driven development) ⭐ Core Skill

Dispatch independent sub-agents for each scheduled task to isolate context from each other.

Each task undergoes a two-stage review after completion: specification compliance review + code quality review.

If the review fails, it will be sent back and redone.

③ test-driven-development(TDD)

Enforce following the “red light-green light-refactor” cycle and write tests first and then code.

④ requesting-code-review (code review)

AI automatically obtains the git diff of the current branch, reviews it file by file, and identifies key issues that prevent advancement.

Installation command:

npx skills add obra/superpowers
# or via Claude Code Official plug-in market
/plugin install superpowers@claude-plugins-official

Applicable scenarios: Core modules for medium and large-scale functional development that require high-quality delivery.

Things to note: Simple tasks will take an extra 15 minutes to complete the process.

3. Top 2: Karpathy Guidelines

It makes AI think like a top engineer.

GitHub Star: 125,000+

The AI ​​coding rules summarized by Andrej Karpathy (co-founder of OpenAI and former Tesla AI director) were compiled into Skills by the community.

Four core rules:

  1. Think before reading code: Do not start grep or modify directly
  2. Modifications should be precise: not "rewrite this file"
  3. Prioritize simplicity: is not a "heap function"
  4. Always check original requirements

Installation command:

/install forrestchang/andrej-karpathy-skills

Applicable scenarios: New project startup, any scenario that requires AI to understand first before taking action.

The value of this Skill lies in what it changes AI behavior patterns, rather than providing a specific tool.

This is especially effective when starting a new project - Claude will be more inclined to understand first and then start, rather than start refactoring directly with confidence.

4. Top 3: Frontend Design

It says goodbye to cookie-cutter “AI style” UI.

Installs: 277,000+

This is the star skill in Anthropic's official Skills repository. If you have used AI-generated front-end code, you will definitely find that they all look similar—the “AI style” that can be seen at a glance.

Frontend Design specializes in solving this problem. It lets AI generateDistinctive and stylishUI instead of a cookie-cutter template style.

Installation command:

npx skills add anthropics/skills --skill frontend-design

Applicable scenarios: Front-end page generation, UI prototype design.

Things to note: It only provides style constraints and cannot replace professional design drafts.

5. Top 4: Webapp Testing

It lets AI do UI automation testing for you.

source: Anthropic official Skills repository

This is a UI automation testing skill based on Playwright. You just tell the AI ​​"Test this login page" and it will automatically:

  • Start browser
  • Perform actions (click, type, navigate)
  • Verify expected results
  • Generate test report

Installation command:

npx skills add anthropics/skills --skill webapp-testing

Applicable scenarios: Web application regression testing, UI automation verification.

6. Top 5: Security Review

It can automatically scan code security vulnerabilities.

source: Claude Code built-in Skill

Security Review is a skill that comes with Claude Code by default. It automatically runs a semantic vulnerability scan before pushing a PR.

Security issues it can detect include:

  • SQL injection risk
  • XSS cross-site scripting vulnerability
  • Sensitive information leaked
  • Unsafe dependency version
  • Authentication and authorization flaws

Usage:

/security-review

Applicable scenarios: Security check and code security audit before submitting PR.

7. Top 6: MCP Builder

It can build an MCP server like building blocks.

source: Anthropic official Skills repository

MCP (Model Context Protocol) is one of the hottest infrastructures in the field of AI programming in 2026.

MCP Builder allows you to automatically generate a complete MCP server through natural language description.

You only need to say: "Build me an MCP server that can query the weather", and the AI ​​will automatically generate:

  • Server skeleton code
  • Tool definition
  • Configuration files and deployment scripts

Installation command:

npx skills add anthropics/skills --skill mcp-builder

Applicable scenarios: Build MCP server, rapid prototyping.

8. Top 7: Skill Creator

It can create skills by itself.

source: Anthropic official Skills repository

Skill Creator is a “meta-Skill” – it’s here to help youCreate new Skill.

You just need to describe what you want the AI ​​to do, and Skill Creator will automatically generate a complete SKILL.md File, containing:

  • YAML frontmatter(name, description)
  • Standardized instruction structure
  • Optional script and reference file templates

Usage:

/skill-creator

Then follow the prompts to enter the description of the Skill you want to create.

Applicable scenarios: Customize exclusive skills for the team and solidify repetitive workflows.

9. Top 8: GSD (Get Shit Done)

Stop talking nonsense and just do it.

GitHub Star: Rapidly growing

The philosophy of GSD is completely opposite to that of Superpowers. Superpowers lets you think clearly before taking action, GSD lets youDon't think about it, just do it.

This skill is suitable for scenarios where you are very sure about what you want to do and don’t need to confirm it again and again.

It skips brainstorming, skips planning, skips reviewing—and goes straight to execution.

Installation command:

git clone https://github.com/gsd-build/get-shit-done ~/.claude/skills/gsd

Applicable scenarios: Repetitive tasks, clear operating instructions, emergency repairs.

Things to note: Only use it when you are 100% sure about what you want to do, otherwise it is easy to overturn.

10. Top 9: GStack

Coding workflow recommended by YC partners.

GitHub Star: 93,000+

Coding workflow skills recommended by Garry Tan (current CEO of YC). It encodes the "rapid iteration, continuous delivery" concept advocated by Y Combinator into an AI-executable workflow.

Installation command:

npx skills add garrytan/gstack

Applicable scenarios: Startup projects develop products quickly and require frequent iterations.

11. Top 10: Composio

It can connect to 1000+ external applications.

GitHub Star: 65,000+

Composio is a connector skill that allows Claude Code to directly operate 1,000+ external applications such as Jira, Linear, Slack, and GitHub.

Once installed, you can direct the AI ​​to:

  • Create tasks in Jira
  • Send a message in Slack
  • Create a PR in GitHub
  • Update work order status in Linear

Installation command:

git clone https://github.com/ComposioHQ/awesome-claude-skills.git
cd awesome-claude-skills
claude --plugin-dir "$PWD/connect-apps-plugin"

Applicable scenarios: Automated workflows that require AI to interact with external tools.

12. Skill installation and usage quick check

Universal installation method

Method 1: Through npx (the most common)

npx skills add <Warehouse name>

Example:npx skills add obra/superpowers

Method 2: Through the Claude Code plug-in market

/plugin install <skillname>@<Market name>

Example:/plugin install superpowers@claude-plugins-official

Method 3: Manual git clone

git clone <Warehouse address> ~/.claude/skills/<skillname>

Skill storage location

LocationScopeWhether to include version control
~/.claude/skills/Personal overall situationno
.claude/skills/Current projectYes (recommended)
Plug-in market installationManaged by pluginno

best practices: Put project-level Skills into .claude/skills/ And submit it to Git. Team members will automatically obtain the same Skill capabilities after cloning.

More practical projects are on my technical website: susan.net.cn/project

Summarize

Back to the original question:What are the 10 most popular Skills in the programming circle in 2026?

Let me use a table to help you summarize:

RankingSkillcore valueThe most suitable scenario
1SuperpowersEnforce engineering discipline and reduce detoursMedium and large function development
2Karpathy GuidelinesCoding thinking from top engineersNew project started
3Frontend DesignSay goodbye to cookie-cutter AI style UIFront-end page generation
4Webapp TestingPlaywright UI automated testingWeb application regression testing
5Security ReviewAutomatic scanning of code security vulnerabilitiesPre-PR security check
6MCP BuilderNatural language generation MCP serverBuild MCP quickly
7Skill CreatorCreate skills yourselfCustomized team-specific skills
8GSDSkip the process and execute it directlyRepeated tasks with clear instructions
9GStackCoding workflow for YC partnersRapid iteration of entrepreneurial projects
10ComposioConnect to 1000+ external applicationsAI interacts with external tools

Some usage suggestions:

  1. Don't be greedy for too much. Installing 50 Skills is not as good as installing 5 fine ones. 70% of community skills are of substandard quality, so selection is more important than selection.

  2. Start with Superpowers. It is the cornerstone of the entire ecosystem. With it installed, you have complete engineering process capabilities.

  3. Add as needed. When you encounter repetitive work, think about it - "Can this be made into a Skill?" Skill Creator can help you quickly turn your ideas into reusable Skills.

  4. Pay attention to skill quality. A good Skill is precise - the description reads like a routing rule, not a slogan; the core SKILL.md is streamlined, and details are pushed to the references/ directory and loaded on demand.

If you haven’t tried Skills yet,It is recommended to install a Superpowers today and try it..

No need to install them all, just start with brainstorming and test-driven-development These are the two most commonly used to start with.

Experience the feeling of "AI thinks clearly before taking action", and you will find that——Programming can be so refreshing.