[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-77830":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":10,"languages":10,"totalLinesOfCode":10,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":13,"stars7d":13,"stars30d":15,"stars90d":14,"forks30d":14,"starsTrendScore":16,"compositeScore":17,"rankGlobal":10,"rankLanguage":10,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":21,"hasPages":19,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":29,"readmeContent":30,"aiSummary":31,"trendingCount":14,"starSnapshotCount":14,"syncStatus":13,"lastSyncTime":32,"discoverSource":33},77830,"agent-skills-in-practice","elisaterumi-ai\u002Fagent-skills-in-practice","elisaterumi-ai","Learn what AI skills are and how to design, structure, and use them in real-world agent systems.","",null,114,26,2,0,7,6,4.29,"Apache License 2.0",false,"main",true,[23,24,25,26,27,28],"agent-skills","ai","ai-agents","claude-code","llm","machine-learning","2026-06-12 02:03:45","🌍 Language:\n[English](README.md) | [Português](README.pt-BR.md) | [Español](README.es.md)\n\n# AI Skills — A Practical Guide\nFrom prompts to systems — learn how to build reusable AI skills.\n\nSkills are one of the core building blocks of modern AI systems. \nThey allow you to define reusable, task-specific capabilities that an AI system can automatically apply when relevant — without repeating instructions every time.\n\n⭐ If this project helped you, please consider giving it a star — it helps others discover it and supports the work!\n\n## Table of Contents\n\n- [Skills Quick Guide](#skills-quick-guide)\n- [What are Skills](#what-are-skills)\n- [How Skills Work](#how-skills-work)\n- [Skill Structure](#skill-structure)\n- [Where Skills Live](#where-skills-live)\n- [Skills vs Other Customizations](#skills-vs-other-customizations)\n- [When to Use Skills](#when-to-use-skills)\n- [Example Use Cases](#example-use-cases)\n- [Creating Your First Skill](#-creating-your-first-skill)\n- [Best Practices](#best-practices)\n- [Common Mistakes](#common-mistakes)\n- [Advanced Concepts](#advanced-concepts)\n- [Why Skills Matter](#why-skills-matter)\n- [Resources](#resources)\n- [Final Thought](#final-thought)\n- [Contributing](#-contributing)\n- [Connect with me](#-connect-with-me)\n\n## Skills Quick Guide\n\nA quick visual reference to understand how skills work — from structure to activation.\n\n\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Felisaterumi-ai\u002Fagent-skills-in-practice\u002Fmain\u002Fimg\u002Fskills-guide-en.png\" width=\"600\"\u002F>\n\n> 💡 Use this as a quick reference while building your own skills.\n\n\n## What are Skills?\n\nA **skill** is a structured set of instructions that teaches an AI system how to perform a specific task.\n\nIn Claude Code, a skill is:\n\n- A directory  \n- Containing a `SKILL.md` file  \n- With metadata and instructions  \n\nOnce defined, the system can discover and use the skill automatically.\n\n> Instead of repeating prompts, you encode behavior once — and reuse it.\n\n## How Skills Work\n\nSkills follow a **semantic matching workflow**:\n\n1. The system loads only skill **names and descriptions**\n2. A user request is received\n3. The system compares the request with available skill descriptions\n4. Matching skills are selected\n5. The full skill content is loaded and executed\n\nThis means:\n\n- Skills are **loaded on demand**\n- They do **not clutter context unnecessarily**\n- They activate **only when relevant**\n\n\n## Skill Structure\n\nEach skill lives inside a directory and must contain a `SKILL.md` file.\n\n### Example:\n\n```yaml\n---\nname: pr-review\ndescription: Reviews pull requests for code quality. Use when reviewing PRs or checking code changes.\n---\n```\n\nBelow the frontmatter, you define the instructions:\n\n```yaml\nWhen reviewing a PR:\n\n1. Analyze code changes\n2. Check for best practices\n3. Suggest improvements\n4. Format output as structured feedback\n```\n\n## Where Skills Live\n\nSkills can be scoped at different levels:\n\n### Personal Skills\n- Location: `~\u002F.claude\u002Fskills`\n- Available across all your projects\n- Ideal for personal workflows\n\n### Project Skills\n- Location: `.claude\u002Fskills` (inside a repository)\n- Shared via version control\n- Used for team standards\n\n## Skills vs Other Customizations\n\n| Feature           | Behavior                | Use Case                |\n|-------------------|------------------------|--------------------------|\n| **Skills**        | Load on demand         | Task-specific expertise |\n| **CLAUDE.md**     | Always loaded          | Global project rules    |\n| **Slash Commands**| Manual invocation      | Explicit actions        |\n\n👉 Skills are **automatic and contextual**\n\n## When to Use Skills\n\nUse a skill when:\n\n- You repeat the same instructions frequently\n- You need consistent output formats\n- You want reusable workflows\n- You want to share knowledge across a team\n\n**Rule of thumb:**\n> If you explain the same thing more than once, it should be a skill.\n\n## Example Use Cases\n\n- Code review standards  \n- Commit message formatting  \n- Documentation templates  \n- Debugging checklists  \n- Data transformation workflows  \n- Clinical text anonymization  \n\n👉 Explore ready-to-use skills: [examples](examples)\n\n> Some examples are inspired by existing open-source skill repositories.\n\n\n## 🚀 Creating Your First Skill\n\n### Step 1 — Create directory\n\n```bash\nmkdir -p ~\u002F.claude\u002Fskills\u002Fmy-skill\n```\n\n### Step 2 — Create `SKILL.md` (or use the template)\n\nCreate a `SKILL.md` file inside your skill directory, or copy from the [provided template](templates\u002Fskill-template.md).\n\n```yaml\n---\nname: my-skill\ndescription: Describe what the skill does and when to use it.\n---\n```\n\n### Step 3 — Add instructions\n\nDefine:\n\n- Steps\n- Rules\n- Output format\n\n## Best Practices\n\n- Write **clear and specific descriptions** (this controls activation)\n- Keep instructions **structured and explicit**\n- Define **expected output formats**\n- Avoid overly generic skills\n- Keep `SKILL.md` concise (use references if needed)\n\n## Common Mistakes\n\n- Treating skills as simple prompts  \n- Writing vague descriptions (skill won’t trigger)  \n- Overloading a single skill with multiple responsibilities  \n- Not defining output structure  \n\n\n## Advanced Concepts\n\n### Metadata Fields\n\n- `name` (required)\n- `description` (required)\n- `allowed-tools` (optional)\n- `model` (optional)\n\n### Progressive Disclosure\n\n- Keep core logic in `SKILL.md`\n- Move large content to:\n  - `references\u002F`\n  - `scripts\u002F`\n  - `assets\u002F`\n\n## Why Skills Matter\n\nSkills transform how we build with AI:\n\n- From **prompts → reusable systems**\n- From **manual instructions → automatic behavior**\n- From **experimentation → production-ready workflows**\n\nThey are a key building block for:\n\n- AI agents  \n- Automation pipelines  \n- Scalable AI systems  \n\n\n\n## Resources\n\nCurated links, documentation, and example repositories:\n\n👉 [Explore resources](docs\u002Fresources.md)\n\n## Final Thought\n\nSkills shift the mindset from:\n\n> “What should I ask the model?”\n\nTo:\n\n> “What capability should this system have?”\n\n## 🤝 Contributing\n\nContributions are welcome!\n\nWays to contribute:\n\n- Add new skills or use cases\n- Improve documentation and examples\n- Report issues or suggest improvements\n\nFeel free to open an issue or submit a pull request 🚀\n\n## 🔗 Connect with me\n\nI share practical insights about AI, agents, and real-world applications:\n\n- LinkedIn (Profile): https:\u002F\u002Fwww.linkedin.com\u002Fin\u002Felisa-terumi  \n- LinkedIn (Page): https:\u002F\u002Fwww.linkedin.com\u002Fcompany\u002Fexploring-artificial-intelligence  \n- Newsletter: https:\u002F\u002Fexploringartificialintelligence.substack.com\u002F  \n- Medium: https:\u002F\u002Fmedium.com\u002F@elisa-terumi\n\n## Star History\n\n\u003Ca href=\"https:\u002F\u002Fwww.star-history.com\u002F?repos=elisaterumi-ai%2Fagent-skills-in-practice&type=date&legend=top-left\">\n \u003Cpicture>\n   \u003Csource media=\"(prefers-color-scheme: dark)\" srcset=\"https:\u002F\u002Fapi.star-history.com\u002Fchart?repos=elisaterumi-ai\u002Fagent-skills-in-practice&type=date&theme=dark&legend=top-left\" \u002F>\n   \u003Csource media=\"(prefers-color-scheme: light)\" srcset=\"https:\u002F\u002Fapi.star-history.com\u002Fchart?repos=elisaterumi-ai\u002Fagent-skills-in-practice&type=date&legend=top-left\" \u002F>\n   \u003Cimg alt=\"Star History Chart\" src=\"https:\u002F\u002Fapi.star-history.com\u002Fchart?repos=elisaterumi-ai\u002Fagent-skills-in-practice&type=date&legend=top-left\" \u002F>\n \u003C\u002Fpicture>\n\u003C\u002Fa>\n","该项目旨在教授如何设计、构建和在实际系统中应用AI技能。核心功能包括通过定义可复用的任务特定能力，使得AI系统能够自动识别并执行相关任务，而无需每次重复指令。技术特点上，项目强调了技能的结构化定义与语义匹配工作流，确保只有当请求与技能描述相匹配时才加载和激活相应技能，从而避免不必要的上下文混乱。适合场景包括但不限于需要提高AI助手效率、简化人机交互过程以及增强自动化处理能力的各种应用场景。","2026-06-11 03:56:10","CREATED_QUERY"]