Skills explained: How Skills compares to prompts, Projects, MCP, and subagents技能解析技能与提示、项目、MCP及子代理的对比关系Skills are an increasingly powerful tool for creating custom AI workflows and agents, but where do they fit in the Claude stack? We explain what tool to use when - and how they all work together.技能是创建自定义AI工作流程和代理日益强大的工具但它们在Claude技术栈中处于什么位置我们将解释何时使用何种工具——以及它们如何协同工作。November 13, 2025https://claude.com/blog/skills-explainedSince introducing Skills, theres been interest in understanding how the various components of Claudes agentic ecosystem work together.Whether youre building sophisticated workflows in Claude Code, creating enterprise solutions with the API, or maximizing your productivity on Claude.ai, knowing which tool to reach for—and when—can transform how you work with Claude.This guide breaks down each building block, explains when to use what, and shows you how to combine them for powerful agentic workflows.自从推出技能功能以来人们一直对了解Claude智能体生态系统中各组件如何协同工作很感兴趣。无论您是在Claude代码中构建复杂工作流通过API创建企业解决方案还是在Claude.ai上最大化工作效率了解何时使用何种工具都能彻底改变您与Claude的协作方式。本指南将分解每个基础组件说明何时使用何种工具并展示如何将它们组合起来构建强大的智能体工作流。Understanding your agentic building blocksWhat are Skills?理解你的代理构建模块什么是技能https://www.youtube.com/watch?vIoqpBKrNaZIAgent Skills: Specialized capabilities you can customizeSkills are folders containing instructions, scripts, and resources that Claude discovers and loads dynamically when relevant to a task. Think of them as specialized training manuals that give Claude expertise in specific domains—from working with Excel spreadsheets to following your organizations brand guidelines.How Skills work:When Claude encounters a task, it scans available Skills to find relevant matches. Skills use progressive disclosure: metadata loads first (~100 tokens), providing just enough information for Claude to know when a Skill is relevant. Full instructions load when needed (5k tokens), and bundled files or scripts load only as required.When to use Skills:Choose Skills when you need Claude to perform specialized tasks consistently and efficiently. Theyre ideal for:Organizational workflows: Brand guidelines, compliance procedures, document templatesDomain expertise:Excel formulas, PDF manipulation, data analysisPersonal preferences:Note-taking systems, coding patterns, research methods技能是包含指令、脚本和资源的文件夹当与任务相关时克劳德会动态发现并加载它们。可以将它们视为专业的培训手册赋予克劳德特定领域的专业知识——从处理Excel电子表格到遵循您组织的品牌指南。技能如何工作当克劳德遇到任务时它会扫描可用技能以查找相关匹配项。技能采用渐进式披露方式首先加载元数据约100个标记仅提供足够的信息让克劳德知道何时技能相关。需要时加载完整指令少于5000个标记捆绑的文件或脚本仅在需要时加载。何时使用技能当您需要克劳德一致且高效地执行专门任务时请选择技能。它们非常适合组织工作流程品牌指南、合规程序、文档模板 领域专业知识Excel公式、PDF操作、数据分析 个人偏好笔记系统、编码模式、研究方法Example:Create a brand guidelines Skill that includes your companys color palette, typography rules, and layout specifications. When Claude creates presentations or documents, it automatically applies these standards without you needing to explain them each time.Learn more about Skills and check out our growing Skills library.What are prompts?https://www.youtube.com/watch?vysPbXH0LpIEPrompts are the instructions you provide to Claude in natural language during a conversation. Theyre ephemeral, conversational, and reactive—you provide context and direction in the moment.When to use prompts:Use prompts for:One-off requests: Summarize this articleConversational refinement: Make that tone more professionalImmediate context: Analyze this data and identify trendsAd-hoc instructions: Format this as a bulleted listExample:Please conduct a comprehensive security review of this code. Im looking for:1. Common vulnerabilities including:Injection flaws (SQL, command, XSS, etc.)Authentication and authorization issuesSensitive data exposureSecurity misconfigurationsBroken access controlCryptographic failuresInput validation problemsError handling and logging issues2. For each issue you find, please provide:Severity level (Critical/High/Medium/Low)Location in the code (line numbers or function names)Explanation of why its a security risk and how it could be exploitedSpecific fix recommendation with code examples where possibleBest practice guidance to prevent similar issues3. Code context: [Describe what the code does, the language/framework, and the environment it runs in - e.g., This is a Node.js REST API that handles user authentication and processes payment data]4. Additional considerations:Are there any OWASP Top 10 vulnerabilities present?Does the code follow security best practices for [specific framework/language]?Are there any dependencies with known vulnerabilities?Please prioritize findings by severity and potential impact.Pro-tip:Prompts are your primary way of interacting with Claude, but they dont persist across conversations. For repeated workflows or specialized knowledge, consider capturing prompts as Skills or project instructions.When to use a Skill instead:If you find yourself typing the same prompt repeatedly across multiple conversations, its time to create a Skill. Transform recurring instructions like review this code for security vulnerabilities using OWASP standards or format this analysis with executive summary, key findings, and recommendations into Skills. This saves you from re-explaining procedures each time and ensures consistent execution.Check out our prompt library, prompting best practices, or our smart prompt maker to get started.What are Projects?Available on all paid Claude plans, Projects are self-contained workspaces with their own chat histories and knowledge bases. Each project includes a 200K context window where you can upload documents, provide context, and set custom instructions that apply to all conversations within that project.How Projects work:Everything you upload to a projects knowledge base becomes available across all chats within that project. Claude automatically uses this context to provide more informed, relevant responses. When your project knowledge approaches context limits, Claude seamlessly enables Retrieval Augmented Generation (RAG) mode to expand capacity by up to 10x.When to use Projects:Choose Projects when you need:Persistent context:Background knowledge that should inform every conversationWorkspace organization:Separate contexts for different initiativesTeam collaboration:Shared knowledge and conversation history (on Team and Enterprise plans)Custom instructions:Project-specific tone, perspective, or approachExample:Create a Q4 Product Launch project containing market research, competitor analysis, and product specifications. Every chat in this project has access to this knowledge without you needing to re-upload or re-explain the context.When to use a Skill instead:Projects give Claude persistent context for a specific body of work—your companys codebase, a research initiative, an ongoing client engagement. Skills teach Claude how to do something. A Project might contain all the background on your product launch, while a skill could teach Claude your teams writing standards or code review process. If you find yourself copying the same instructions across multiple Projects, thats a signal to create a skill instead.Learn more about Projects.What are subagents?Subagents are specialized AI assistants with their own context windows, custom system prompts, and specific tool permissions. Available in Claude Code and the Claude Agent SDK, subagents handle discrete tasks independently and return results to the main agent.How subagents work:Each subagent operates with its own configuration—you define what it does, how it approaches problems, and which tools it can access. Claude automatically delegates tasks to appropriate subagents based on their descriptions, or you can explicitly request a specific subagent.When to use subagents:Use subagents for:Task specialization:Code review, test generation, security auditsContext management:Keep the main conversation focused while offloading specialized workParallel processing:Multiple subagents can work on different aspects simultaneouslyTool restriction:Limit specific subagents to safe operations (e.g., read-only access)Example:Create a code-reviewer subagent with access to Read, Grep, and Glob tools but not Write or Edit. When you modify code, Claude automatically delegates to this subagent for quality and security review without risking unintended code changes.When to use a Skill instead:If multiple agents or conversations need the same expertise—like security review procedures or data analysis methods—create a Skill rather than building that knowledge into individual subagents. Skills are portable and reusable, while subagents are purpose-built for specific workflows. Use Skills to teach expertise that any agent can apply; use subagents when you need independent task execution with specific tool permissions and context isolation.Learn more about subagents.What is MCP?The Model Context Protocol (MCP) is an open standard for connecting AI assistants to external systems where data lives—content repositories, business tools, databases, and development environments.How MCP works:MCP provides a standardized way to connect Claude to your tools and data sources. Instead of building custom integrations for each data source, you build against a single protocol. MCP servers expose data and capabilities; MCP clients (like Claude) connect to these servers.When to use MCP:Choose MCP when you need Claude to:Access external data: Google Drive, Slack, GitHub, databasesUse business tools: CRM systems, project management platformsConnect to development environments: Local files, IDEs, version controlIntegrate with custom systems: Your proprietary tools and data sourcesExample:Connect Claude to your companys Google Drive via MCP. Now Claude can search documents, read files, and reference internal knowledge without manual uploads—the connection persists and updates automatically.When to use a Skill instead:MCP connects Claude to data; Skills teach Claude what to do with that data. If youre explaininghowto use a tool or follow procedures—like when querying our database, always filter by date range first or format Excel reports with these specific formulas—thats a Skill. If you need Claude toaccessthe database or Excel files in the first place, thats MCP. Use both together: MCP for connectivity, Skills for procedural knowledge.Learn more about MCP and check out documentation on how to build an MCP server.How they work togetherThe real power emerges when you combine these building blocks. Each serves a distinct purpose, and together they create sophisticated agentic workflows.Comparison: choosing the right toolFeatureSkillsPromptsProjectsSubagentsMCPWhat it providesProcedural knowledgeMoment-to-moment instructionsBackground knowledgeTask delegationTool connectivityPersistenceAcross conversationsSingle conversationWithin projectAcross sessionsContinuous connectionContainsInstructions code assetsNatural languageDocuments contextFull agent logicTool definitionsWhen it loadsDynamically, as neededEach turnAlways in projectWhen invokedAlways availableCan include codeYesNoNoYesYesBest forSpecialized expertiseQuick requestsCentralized contextSpecialized tasksData accessExample agentic workflow: research agentLets build a comprehensive research agent that combines multiple building blocks. This example shows how to assemble and activate an agent for competitive analysis.Step 1: Set up your ProjectCreate a Competitive Intelligence project and upload:Industry reports and market analysesCompetitor product documentationCustomer feedback from your CRMPrevious research summariesAdd project instructions:Analyze competitors through the lens of our product strategy. Focus on differentiation opportunities and emerging market trends. Present findings with specific evidence and actionable recommendations.Step 2: Connect data sources via MCPEnable MCP servers for:Google Drive (to access shared research documents)GitHub (to review competitor open-source repositories)Web search (for real-time market information)Step 3: Create specialized SkillsCreate a competitive-analysis skill:# My Company GDrive Navigation Skill ## Overview Optimized search and retrieval strategy for Meridian Techs Google Drive structure. Use this skill to efficiently locate internal documents, research, and strategic materials. ## Drive Organization **Top-level structure:** - /Strategy Planning/ - OKRs, quarterly plans, board decks - /Product/ - PRDs, roadmaps, technical specs - /Research/ - Market research, competitive intel, user studies - /Sales Marketing/ - Case studies, pitch decks, campaign materials - /Customer Success/ - Implementation guides, success metrics - /Company Ops/ - Policies, org charts, team directories **Naming conventions:** - Format: YYYY-MM-DD_DocumentName_vX - Final versions marked with _FINAL - Drafts include _DRAFT or _WIP ## Search Best Practices 1. **Start broad, then filter** - Use folder context keywords 2. **Target document owners** - Sales materials from Sales/, not root 3. **Check recency** - Prioritize documents from last 6 months for current strategy 4. **Look for source of truth** - Files with _FINAL, _APPROVED, or in /Archives/Official/ ## Research Agent Workflow 1. Identify topic category (product, market, customer) 2. Search relevant folder with targeted keywords 3. Retrieve 3-5 most recent/relevant documents 4. Cross-reference with /Strategy Planning/ for context 5. Cite sources with file names and datesStep 4: Configure subagents (Claude Code/SDK only)Create specialized subagents:market-researchersubagent:name: market-researcher description: Research market trends, industry reports, and competitive landscape data. Use proactively for competitive analysis. tools: Read, Grep, Web-search --- You are a market research analyst specializing in competitive intelligence. When researching: 1. Identify authoritative sources (Gartner, Forrester, industry reports) 2. Gather quantitative data (market share, growth rates, funding) 3. Analyze qualitative insights (analyst opinions, customer reviews) 4. Synthesize trends and patterns Present findings with citations and confidence levels.technical-analystsubagent:name: technical-analyst description: Analyze technical architecture, implementation approaches, and engineering decisions. Use for technical competitive analysis. tools: Read, Bash, Grep --- You are a technical architect analyzing competitor technology choices. When analyzing: 1. Review public repositories and technical documentation 2. Assess architecture patterns and technology stack 3. Evaluate scalability and performance approaches 4. Identify technical strengths and limitations Focus on actionable technical insights that inform our product decisions.Step 5: Activate your research agentNow when you ask Claude: Analyze how our top three competitors are positioning their new AI features and identify gaps we can exploitHeres what happens:Project context loads: Claude accesses your uploaded research documents and follows project instructionsMCP connections activate: Claude searches your Google Drive for recent competitor briefs and pulls GitHub dataSkills engage: The competitive-analysis Skill provides the analytical frameworkSubagents execute(in Claude Code): The market-researcher gathers industry data while the technical-analyst reviews technical implementationsPrompts refine: You provide conversational guidance: Focus especially on enterprise customers in healthcareThe result:A comprehensive competitive analysis that draws from multiple data sources, follows your analytical framework, leverages specialized expertise, and maintains context throughout your research project.Common questionsHow do Skills work?Skills use progressive disclosure to keep Claude efficient. When working on tasks, Claude first scans Skill metadata (descriptions and summaries) to identify relevant matches. If a Skill matches, Claude loads the full instructions. Finally, if the Skill includes executable code or reference files, those load only when needed.This architecture means you can have many Skills available without overwhelming Claudes context window. Claude accesses exactly what it needs, when it needs it.Skills vs. subagents: when to use whatUse Skills when:You want capabilities that any Claude instance can load and use. Skills are like training materials—they make Claude better at specific tasks across all conversations.Use subagents when:You need complete, self-contained agents designed for specific purposes that handle workflows independently. Subagents are like specialized employees with their own context and tool permissions.Use them together when:You want subagents with specialized expertise. For example, a code-review subagent can use Skills for language-specific best practices, combining the independence of a subagent with the portable expertise of Skills.Skills vs. prompts: when to use whatUse prompts when:Youre giving one-time instructions, providing immediate context, or having a conversational back-and-forth. Prompts are reactive and ephemeral.Use Skills when:You have procedures or expertise that youll need repeatedly. Skills are proactive—Claude knows when to apply them—and persistent across conversations.Use them together:Prompts and Skills complement each other naturally. Use Skills to provide foundational expertise, then use prompts to provide specific context and refinement for each task.Skills vs. Projects: when to use whatUse Projects when:You need background knowledge and context that should inform all conversations about a specific initiative. Projects provide static reference material thats always loaded.Use Skills when:You need procedural knowledge and executable code that activates only when relevant. Skills provide dynamic expertise that loads on-demand, saving your context window.Use them together when:You want both persistent context and specialized capabilities. For example, a Product Development project containing product specs and user research, combined with Skills for creating technical documentation and analyzing user feedback data.Key difference:Projects say heres what you need to know. Skills say heres how to do things. Projects provide a knowledge base you work within. Skills provide capabilities that work everywhere—any conversation, any project.Can subagents use Skills?Yes. In Claude Code and the Agent SDK, subagents can access and use Skills just like the main agent. This creates powerful combinations where specialized subagents leverage portable expertise.For example, your python-developer subagent can use the pandas-analysis Skill to perform data transformations following your teams conventions, while your documentation-writer subagent uses the technical-writing skill to format API documentation consistently.Getting startedReady to build with Skills? Heres how to start:Claude.aiusers:Enable Skills in Settings → FeaturesCreate your first project at claude.ai/projectsTry combining project knowledge with Skills for your next analysis taskAPI developers:Explore the Skills endpoint in documentationCheck out our skills cookbookClaude Code users:Install Skills via plugin marketplacesCheck out our skills cookbook--