mcpbeat Sign in

Claude Skills

The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.

Every Agent Skill we could find on GitHub, deduplicated by content. 86 696 files from 2 113 authors, of which 68 413 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.

68 413
unique skills
out of 86 696 files found on GitHub
18 283
are copies
same content, someone else's repository
1 744
tokens, median
what a typical skill costs you in context
8 409
name collisions
two skills with one name cannot sit side by side

64 201–64 260 of 68 413

page 1071 of 1 141
Shopify Storefront Graphql
by Shopify
vendor
778 tokens scripts
Skill Interview Builder
by irenerachel

| 通过分步访谈引导用户理清需求,最终产出完整的Skill文件包(含SKILL.md、参考文档、示例文件等), 并打包为可直接使用的压缩包。 当用户说"我想通过访谈新建Skill"、"用访谈方式做一个Skill"、"访谈建Skill"、 "通过访谈帮我生成Skill"、"访谈式创建Skill"、"我想访谈做一个XX的技能"时触发。 触发关键词必须包含"访谈"二字,不含"访谈"的Skill创建请求不由本Skill处理。 不用于已有完整SKILL.md只需小改的情况,也不用于一次性提示词请求。

2.6k tokens zh
Interactive Learning
by geekjourneyx

| 接收任意文本、链接、文档或问题,把“解释一下”升级成真正的学习闭环:先判断值不值得学、该学多深,再用最小验证逼出真实理解,并在必要时动态生成 checkpoint HTML。只要用户明显想“学懂”“吃透”“带我学”“帮我验证我是不是真懂了”,而不是只要摘要、改写或普通解释,就应主动使用本技能。也适用于用户想导出学习卡片、checkpoint 页面、复盘页、学习状态页的场景。

1.5k tokens zh
Satori
by MetcalfSolutions

Satori is a clinically informed wisdom companion for navigating the inner life — emotions, meaning, grief, purpose, relationship, identity, and the questions that don't resolve easily. Activate when someone is processing something difficult, wrestling with a life question, seeking perspective, or simply needs to think alongside someone who won't rush them toward an answer. Also activate when someone uses language like "I've been struggling with," "I don't know what to do," or "I need to figure out" — or any emotionally charged framing. When in doubt, activate. Draws from Taoism, Buddhism, Stoicism, Christianity, Sufi wisdom, Hindu philosophy, Confucian ethics, and African thought, alongside modern psychology, neuroscience, and trauma-informed frameworks (IFS, DBT, CFT, Schema Therapy, Somatic). Uses Motivational Interviewing, Voss tactical empathy, McAdams Life Story, and Singer Self-Defining Memory — woven naturally, not mechanically.

1.2k tokens scripts
Ducksearch
by liangdabiao

使用 DuckDuckGo 进行网页搜索和内容提取的命令行工具。当用户需要搜索网络信息、查找资料、获取网页内容时使用此 skill。触发场景包括:(1) 搜索网络内容 (2) 获取网页文本 (3) 使用 DuckDuckGo 搜索 (4) 抓取网页内容 (5) 配置 MCP 搜索服务器。

359 tokens zh
Research Brightdata
by liangdabiao

This skill should be used when the user asks to "research web data", "scrape websites", "extract web data", "perform market research", "analyze competitors", "monitor prices", "collect product information", "search and analyze web content", or mentions Bright Data MCP, web scraping, web data extraction, or automated research. Provides comprehensive web research workflows using Bright Data MCP tools including search, scraping, extraction, and browser automation capabilities.

2.7k tokens
Tagore
by apurvrdx1

| Write or rewrite prose so it sounds like a human wrote it — not a frontier model. Named in homage to Rabindranath Tagore, whose prose carried what a 29-pattern catalog of AI tells (from humanizer) plus an 8-rule operating system with an 8-dimension scoring gate (extending stop-slop). Use when emails. Detects and removes inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary, passive voice, negative parallelisms, filler phrases, inanimate-verb constructions, narrator-from-a-distance voice, and metronomic specificity, restraint, varied rhythm, and trust in the reader.

9.1k tokens scripts
CrowdStrike Fusion Workflow Builder
by eth0izzle

> Create, validate, import, execute, and export CrowdStrike Falcon Fusion SOAR action IDs BEFORE writing any YAML. NEVER write PLACEHOLDER values for action IDs — resolve every ID via the live API first. Templates and example files in this repo contain PLACEHOLDER markers that are structural guides only — do NOT copy them into output YAML. For plugin config_id values, ask the user. Use this skill when asked to create a CrowdStrike workflow, Fusion workflow, Falcon Fusion automation, SOAR playbook, build a workflow for CrowdStrike, automate CrowdStrike actions, or anything involving CrowdStrike Fusion SOAR.

4.2k tokens scripts
Falcon Next Gen SIEM Lookup Files
by eth0izzle

> Create, list, download, update, and delete CrowdStrike Falcon Next-Gen SIEM lookup files. Upload CSV or JSON files for use with the match() function in CrowdStrike Query Language (CQL) queries. Use this skill when asked to manage lookup files, upload CSV data to CrowdStrike, create reference tables for SIEM queries, or work with Falcon Next-Gen SIEM lookup file operations.

1.8k tokens scripts
Council Implement
by SamJHudson01

Execute a Carmack Council plan task by task. Use when explicitly asked to implement a plan, do a "council implement", "carmack implement", "council build", or invoke /council-implement. Reads the output of /council-plan and builds each task sequentially, loading the relevant expert's reference document per task. Verifies after each task. Produces an implementation log for /council-review. Stack: Next.js App Router / React / TypeScript / tRPC / Prisma / Neon / Clerk.

2.6k tokens
Spec Writer
by SamJHudson01

Generate structured software specifications for features, bug fixes, and products. Use when the user wants to create a spec, PRD, feature brief, requirements document, or when starting any new implementation that needs a specification first. Invoke via /spec-writer or when the user says "write a spec", "spec this out", "create a spec", "I need a spec for...", or describes a feature they want to build. Produces adaptive-complexity specs with Job Stories, Gherkin acceptance criteria, and three-tier boundaries. Output is a markdown file ready for agent execution or human review.

3.6k tokens
Council Plan
by SamJHudson01

Architect a feature with the Carmack Council before writing code. Use when explicitly asked to plan a feature, do a "council plan", "carmack plan", or invoke /council-plan. Carmack's philosophy chairs a council of domain experts — Troy Hunt (security), Martin Fowler (refactoring), Kent C. Dodds (frontend), Matteo Collina (Node.js), Brandur Leach (Postgres), Vercel Performance, Simon Willison (LLM pipeline), Karri Saarinen (UI quality), Vitaly Friedman (UX quality). Interactive feature discovery followed by parallel subagent dispatch. Produces a sequenced, attributed implementation plan with no code. Stack: Next.js App Router / React / TypeScript / tRPC / Prisma / Neon / Clerk.

7.3k tokens
Council Review
by SamJHudson01

Perform a rigorous Carmack Council code review. Use when explicitly asked to review code, do a "council review", "carmack review", or invoke /council-review. Carmack's philosophy chairs a council of domain experts — Troy Hunt (security), Martin Fowler (refactoring), Kent C. Dodds (frontend), Matteo Collina (Node.js), Brandur Leach (Postgres), Vercel Performance, Simon Willison (LLM pipelines), Karri Saarinen (UI quality), Vitaly Friedman (UX quality), Kent Beck (test quality). Uses parallel subagents for deep, independent review. Produces prioritised P1/P2/P3 findings. Stack: Next.js App Router / React / TypeScript / tRPC / Prisma / Neon / Clerk.

12k tokens
Test Architect
by SamJHudson01

Map testable surfaces, audit existing tests for quality, and write test specifications that prevent AI shortcuts. Use when asked to "audit tests", "specify tests", "test architect", "map test coverage", or invoke /test-architect. Two modes — audit (evaluate existing tests against Beck's principles) and specify (write test specs for new code). Powered by Carmack × Beck quality-testing.md reference doc. Stack: Next.js App Router / TypeScript / Vitest / Cypress / tRPC / Prisma / Neon.

5.7k tokens
Rrr
by dgk-dev

GLM-5 코드 리뷰. /rr의 상위 버전으로 더 강한 모델로 변경사항을 검토하고, Codex가 결과를 다시 검증해 유효한 이슈만 정리한다.

599 tokens
Rr
by dgk-dev

코드 리뷰. 작업 후 변경사항을 Z.AI 모델로 검토하고, Codex가 결과를 다시 검증해 유효한 이슈만 정리한다. /rr 또는 'GLM 리뷰' 요청 시 사용.

672 tokens
Fd
by dgk-dev

Frontend design mode for strong visual direction, UI polish, redesigns, landing pages, and production-grade interface work, including shadcn-based surfaces. Use when the user says `/fd` or explicitly asks for better design, styling, layout, visual quality, or UX polish.

2.2k tokens
Re
by dgk-dev

Explicit extra-research mode. Use only when the user says `/re` or clearly asks for a research-heavy pass before coding. This skill should bias Codex toward more source-checking and justification without replacing its normal orchestration.

547 tokens
Ralph
by dgk-dev

Persistent completion mode. Use when the user explicitly says `/ralph` or clearly wants you to keep iterating until the task is actually finished, repeating implement-verify-fix loops instead of stopping at partial progress.

326 tokens
Cp
by dgk-dev

Local verify-commit-push wrapper for Codex. Use only when the user explicitly says `/cp` or clearly asks to finish the current work by running the relevant local checks, staging intended files explicitly, committing on the current branch, and pushing. This workflow may commit directly to `main` or `master` when that is the repo's chosen operating model.

658 tokens
AI Native Knowledge RAG
by gmaxxxie

| AI Native 产品方法论——RAG与知识系统设计的实操 Skill。 用户提供企业知识场景,Skill 自动执行知识系统设计流程: 资料来源分析 → 清洗与脱敏 → 索引与权限控制 → 检索召回 → 评估与更新 → 输出知识系统方案。 基于《AI Native 产品方法论》第15章。

5.8k tokens zh
AI Native Pm Agent
by gmaxxxie

| AI Native PM Agent — 八书合一的产品全生命周期编排器。 用户提供产品想法或问题线索,Agent 自动路由到 P0-P14 对应阶段执行。 基于8本书的方法论体系,80个可执行Skill。

1.1k tokens
AI Native Direction Framing
by gmaxxxie

| AI Native 产品方法论——方向定界阶段的实操 Skill。 用户提供一个 AI 产品方向或问题线索,Skill 自动执行方向定界流程: 问题真实性判断 → 场景切入分析 → 资料条件审查 → 能力可能性评估 → 价值判断 → 输出 Direction Brief。 基于《AI Native 产品方法论》第05章。

4.7k tokens zh
AI Native Audit Release
by gmaxxxie

| AI Native 产品方法论——审计放行阶段的实操 Skill。 用户提供系统构建方案,Skill 自动执行审计放行流程: 设计证据 → 评估证据 → Shadow 证据 → 放行边界判断 → go/no-go 决策 → 输出放行方案。 基于《AI Native 产品方法论》第17章。

7.9k tokens zh
AI Native Business Model
by gmaxxxie

AI Native 商业模式设计 Skill。基于《AI确定性商业模式》方法论, 帮助用户设计以"确定性溢价"为核心的 AI 商业模式: 避开6种失效模式,选择4种确定性模型,构建可持续的收费逻辑。

6.4k tokens zh
AI Native Context Engineering
by gmaxxxie

| AI Native 产品方法论——上下文工程的实操 Skill。 用户提供任务场景,Skill 自动执行上下文组织流程: 任务目标识别 → 上下文层选择 → 动态拼装 → 成本与窗口控制 → 结果校验与纠偏 → 输出上下文工程方案。 基于《AI Native 产品方法论》第14章。

6.3k tokens zh
AI Native Agent Skill Design
by gmaxxxie

| AI Native 产品方法论——智能体与技能单元设计的实操 Skill。 用户提供任务场景,Skill 自动执行能力编排设计流程: 任务拆解 → Agent 角色定义 → Skill 拆分 → Tool 映射 → 边界与回退设计 → 输出能力编排方案。 基于《AI Native 产品方法论》第12章。

3.2k tokens zh
AI Native Experiment Engine
by gmaxxxie

| AI Native 产品方法论——试验展开阶段的实操 Skill。 用户提供 Direction Brief 或问题方向,Skill 自动执行试验展开流程: 资料准备 → 能力实验 → 产品实验 → 商业实验 → 评估与失败分析 → 输出实验结论报告。 覆盖《AI Native 产品方法论》第06-10章。

5.2k tokens zh
AI Native Memory System
by gmaxxxie

| AI Native 产品方法论——记忆系统设计的实操 Skill。 用户提供产品场景,Skill 自动执行记忆系统设计流程: 记忆需求分析 → 记忆分类设计 → 权限与时效策略 → 存储与索引 → 人工修正与再沉淀 → 输出记忆系统方案。 基于《AI Native 产品方法论》第13章。

2.4k tokens zh
AI Native Marketing Growth
by gmaxxxie

| AI Native 营销与增长策略 Skill。基于《AI Native 营销与增长》方法论, 帮助用户构建以 AI 为第一性原理的增长系统:从工具思维升级为系统思维, 建立 AI Native 增长飞轮,实现数据-模型-反馈的复利增长。

7.3k tokens zh
AI Native Pm Agent
by gmaxxxie

AI Native 产品经理 Agent — — 五书合一的产品工作流编排器。 用户提供一个产品想法或问题线索,Agent 自动编排七个阶段: P1需求发现 → P2方向定界 → P3体验设计 → P4系统构建 → P5商业模式 → P6增长策略 → P7审计投产。 基于五本书:《AI Native 产品方法论》《AI时代的用户体验》《AI确定性商业模式》《AI Native 营销与增长》《AI rebuild product needs》。

3.2k tokens zh
AI Native Product Needs
by gmaxxxie

AI Native 产品需求发现 Skill。基于《AI rebuild product needs》方法论, 帮助用户在 AI Agent 时代重新定义需求:区分真实需求与伪需求, 以"场景为入口、处境为本体",用行为证据替代表达性需求, 最终输出一份经过验证的 AI Native 需求简报。

7.1k tokens zh
AI Native Ux Design
by gmaxxxie

AI Native 产品方法论——AI Native 用户体验设计的实操 Skill。 用户提供产品场景,Skill 自动执行 UX 设计流程: 任务分析 → 人机分工设计 → 状态可见性 → 纠偏机制 → 信任设计 → 反馈沉淀 → 输出 UX 方案。 基于《AI Native 产品方法论》第16章。 '

6.7k tokens zh
AI Native System Building
by gmaxxxie

| AI Native 产品方法论——系统构建阶段的总论 Skill。 用户提供实验结论报告,Skill 自动执行系统构建流程: 实验证据输入 → 系统边界定义 → 能力模块设计 → 治理与观测补齐 → 输出系统构建方案。 基于《AI Native 产品方法论》第11章。

5.7k tokens zh
AI Native Ux Design
by gmaxxxie

| AI Native 产品方法论——AI Native 用户体验设计的实操 Skill。 用户提供产品场景,Skill 自动执行 UX 设计流程: 任务分析 → 人机分工设计 → 状态可见性 → 纠偏机制 → 信任设计 → 反馈沉淀 → 输出 UX 方案。 基于《AI Native 产品方法论》第16章。

5.3k tokens zh
AI Native Production Ops
by gmaxxxie

| AI Native 产品方法论——生产运行与循环回灌的实操 Skill。 用户提供产品已上线或即将上线,Skill 自动执行生产运行流程: 可观测性设计 → 指标体系搭建 → 失败与人工修正识别 → 反馈回灌 → 价值/商业/客户循环 → 输出生产运行方案。 包含价值发现循环(§1)、AIOps案例(§2)、客服案例(§3)、SaaS案例(§4)、产品团队(§5)五个合并子模块。 基于《AI Native 产品方法论》第18-27章。

9.9k tokens zh
Combo Needs To Direction
by gmaxxxie

需求发现 → 方向定界 组合 Skill。从用户痛点出发,系统跑完微需求检测、 真需求验证、四层拆解,输出可直接进入方向定界的 Direction Brief。

3.2k tokens zh
Combo Ux To Audit
by gmaxxxie

UX 设计 → 审计放行 组合 Skill。将 RAX 风险评估、信任度分级、渐进式披露 与审计放行流程打通,输出 UX 审计报告 + 放行建议。

3.3k tokens zh
Combo Business To Growth
by gmaxxxie

商业模式 → 增长策略 组合 Skill。将确定性溢价定价策略与 AI Native 增长飞轮匹配, 输出定价-增长联动策略,确保商业模式和增长方式自洽。

3.4k tokens zh
P0 Needs Orchestrator
by gmaxxxie

需求发现编排器。协调微需求检测、真需求验证、需求拆解、需求考古、好问题生成、Agent 边界设计、多元推荐改写、AI 产品三重平衡八个工具卡, 按顺序执行并输出完整的需求发现报告。自包含——八个工具卡的核心方法已内嵌。 基于《AI rebuild product needs》读者工具箱。

4.9k tokens zh
AI Native Product Needs
by gmaxxxie

AI Native 产品需求发现 Skill。基于《AI rebuild product needs》方法论, 帮助用户在 AI Agent 时代重新定义需求:区分真实需求与伪需求, 以"场景为入口、处境为本体",用行为证据替代表达性需求, 最终输出一份经过验证的 AI Native 需求简报。

4.3k tokens zh
P0c Needs Decomposer
by gmaxxxie

需求四层拆解卡。很多团队一听到用户表达,就直接拆功能。 这张卡强迫团队先把"用户说了什么"和"用户真正被什么困住"分开。 基于《AI rebuild product needs》工具卡。

3.6k tokens zh
P0a Micro Needs Detector
by gmaxxxie

微需求五问检测器。团队天然偏爱"大需求",这个 Skill 把看起来很小、但可能压得很深的问题重新看见。 基于《AI rebuild product needs》工具卡。

3.5k tokens zh
P0b Real Needs Validator
by gmaxxxie

真需求判断五问验车器。AI 时代最危险的不是没有需求,而是伪需求太容易长得像真需求。 基于《AI rebuild product needs》工具卡。

3.8k tokens zh
P0d Needs Archaeologist
by gmaxxxie

需求考古五步法。用户说出来的需求只是冰山一角,真正的需求往往埋在历史决策、组织结构和业务演化中。 这个 Skill 教你如何挖掘深层需求。 基于《AI rebuild product needs》工具卡。

4.4k tokens zh
P0g Diverse Recommendation Rewriter
by gmaxxxie

多元推荐改写清单。AI 时代的推荐系统不再是"猜你喜欢", 而是"帮你发现你还不知道自己会喜欢的"。这个 Skill 提供从单一推荐到多元推荐的改写框架。 基于《AI rebuild product needs》工具卡。

3.4k tokens zh
P0e Good Question Generator
by gmaxxxie

好问题六维观察表。AI 时代产品经理最重要的能力不是给答案,而是提好问题。 这个 Skill 提供六个维度的观察框架,帮你从不同角度发现好问题。 基于《AI rebuild product needs》工具卡。

3.5k tokens zh
P0f Agent Boundary Designer
by gmaxxxie

Agent 边界清单。AI 产品最危险的不是做不到,而是做过头。 这个 Skill 帮你系统性地定义 AI 应该做什么、不应该做什么、什么时候必须停下来等人。 基于《AI rebuild product needs》工具卡。

3.9k tokens zh
P0g Diversity Rewrite Checklist
by gmaxxxie

多元推荐改写清单。当团队想重写推荐系统时,最容易只在原目标函数上加一点随机。 这个 Skill 帮你避免"看起来更多元,底层仍然单一"。 基于《AI rebuild product needs》工具卡。

3.9k tokens zh
AI Native Direction Framing
by gmaxxxie

AI Native 产品方法论——方向定界阶段的实操 Skill。 用户提供一个 AI 产品方向或问题线索,Skill 自动执行方向定界流程: 问题真实性判断 → 场景切入分析 → 资料条件审查 → 能力可能性评估 → 价值判断 → 输出 Direction Brief。 基于《AI Native 产品方法论》第05章。 '

4.4k tokens zh
P0h Triple Balance Assessor
by gmaxxxie

AI 产品三重平衡表。当团队讨论 AI 产品方向时,最容易只看增长和效率。 这个 Skill 帮团队同时看商业、人性、文化三层。 基于《AI rebuild product needs》工具卡。

4.1k tokens zh
P0h AI Product Triple Balance
by gmaxxxie

AI 产品三重平衡表。AI 产品设计需要在商业、人性、技术三个维度之间找到平衡。 这个 Skill 提供系统性的三重平衡评估框架。 基于《AI rebuild product needs》工具卡。

4.1k tokens zh
P10a Value Discovery Loop
by gmaxxxie

AI Native 产品方法论——价值发现循环的实操 Skill。 用户提供已上线或正在验证的 AI 产品,Skill 自动执行价值发现流程: 价值信号识别 → 伪价值排除 → 价值密度评估 → 市场测试设计 → 定价策略联动 → 方向修正建议 → 输出价值发现方案。 基于《AI Native 产品方法论》第22章。 '

5k tokens zh
AI Native Production Ops
by gmaxxxie

AI Native 产品方法论——生产运行与循环回灌的实操 Skill。 用户提供产品已上线或即将上线,Skill 自动执行生产运行流程: 可观测性设计 → 指标体系搭建 → 失败与人工修正识别 → 反馈回灌 → 价值/商业/客户循环 → 输出生产运行方案。 基于《AI Native 产品方法论》第18-27章。 '

2.7k tokens zh
P10b Aiops Case
by gmaxxxie

AI Native 产品方法论——AIOps 行业案例模板 Skill。 用户提供 AIOps 场景和业务背景,Skill 自动执行全链路方法论: 方向定界 → 试验展开 → 系统构建 → 审计放行 → 生产运行,输出完整案例方案。 基于《AI Native 产品方法论》第19章 AIOps 案例,适用于高风险、强约束、重责任场景。 '

4.6k tokens zh
P10c Customer Service Case
by gmaxxxie

AI Native 产品方法论——AI 客服行业案例模板 Skill。 用户提供客服场景和业务背景,Skill 自动执行全链路方法论: 方向定界 → 试验展开 → 系统构建 → 审计放行 → 生产运行 → 客户循环 → 价值发现,输出完整客服产品方案。 基于《AI Native 产品方法论》第20章 AI 客服案例,适用于服务协同、经验复利场景。 '

4.9k tokens zh
P10d Saas Case
by gmaxxxie

AI Native 产品方法论——AI Native SaaS 行业案例模板 Skill。 用户提供 SaaS 产品场景和业务背景,Skill 自动执行全链路方法论: 方向定界 → 试验展开 → 系统构建 → 审计放行 → 生产运行,输出 AI Native 化改造方案。 基于《AI Native 产品方法论》第21章 AI Native SaaS 案例,适用于数据分析、业务洞察、决策支持场景。 '

4.9k tokens zh
P11 Product Team
by gmaxxxie

AI Native 产品方法论——AI Native 产品团队设计的实操 Skill。 用户提供产品阶段和团队现状,Skill 自动执行团队设计: 角色梳理 → 人机分工设计 → 能力缺口识别 → 团队协作流程 → 能力建设路径 → 输出 AI Native 团队方案。 基于《AI Native 产品方法论》第25章 AI Native 产品团队。 '

5.3k tokens zh
P12 Contemplation Orchestrator
by gmaxxxie

《观照》全书编排器——根据用户输入的问题类型,路由到对应的章节Skill,串联全链路决策修正流程

3.9k tokens zh
P12a Contemplation Prerequisite Check
by gmaxxxie

检查方法使用的前提是否成立——识别情境变化、验证假设前提、重配适用方法

4.2k tokens zh

Claude Skills — questions

Answers built from the skills we actually parsed.

What is a Claude Skill?
A folder with a SKILL.md file: instructions that teach an agent to do one thing well, optionally with scripts and reference files alongside. The format is open and called Agent Skills — Claude Code, Codex and other agents read the same files. It is not a program you run; it is knowledge the agent loads when the task calls for it.
How is a skill different from an MCP server?
A server gives the agent new abilities — it connects to something and exposes tools. A skill gives the agent knowledge: how to use what it already has. They combine, and often literally: 13 362 of the skills here declare which MCP servers they need to work.
Why are there fewer skills here than in other catalogues?
Because we deduplicate by content. Of 86 696 files found on GitHub, 68 413 are unique — the rest is the same skill copied into someone else's repository, word for word. Catalogues that count files rather than skills show every copy as a separate entry.
What does the token count mean?
A skill is loaded in stages. Its name and description stay in the model's context all the time; the body of SKILL.md is read when the skill is used; other files in the folder, such as references, scripts and fonts, are opened only when the skill asks for them. The number we show is SKILL.md, the part every use pays for. A big folder is not a big cost: one official skill ships 83 font files, and none of them ever enters the context.
How do I install a skill?
Copy the skill folder into ~/.claude/skills for personal use, or into .claude/skills inside a project. The agent picks it up by the name in the SKILL.md header — which is worth checking: 8 409 skills here share a name with another skill, and two of them cannot sit side by side.