# Bigger than MCP? The Claude Agent Skills is AWESOME

Source: https://www.youtube.com/watch?v=1WImBwiA7RA
Recap page: https://rapidrecap.app/video/1WImBwiA7RA
Generated: 2025-10-18T05:33:08.458+00:00

---
## Quick Overview

Anthropic's new Agent Skills feature provides a structured, token-efficient, and composable way to equip Claude agents with domain-specific procedural knowledge and tools, potentially surpassing the utility of traditional Manifest Component Protocols (MCPs) by embedding necessary context directly within skill directories.

**Key Points:**
- Agent Skills are introduced by Anthropic as a new pattern for building specialized agents using organized folders of instructions, scripts, and resources.
- A Skill is defined as a combination of a prompt (the SKILL.md file) and optional tools/assets (files within a 'templates' folder), which can also be prompt-only.
- Skills offer significant token efficiency improvements over MCPs, as demonstrated by reducing tool context from 4.2k tokens down to 70 tokens by only loading relevant files.
- The video showcases creating custom skills based on an existing codebase, such as a 'slack-gif-creator' skill using predefined Python functions and a 'frontend' skill detailing best practices for UI development.
- The structure allows agents to dynamically load context-specific information, like brand guidelines or component implementation logic, only when needed, leading to better performance and composability.
- The speaker suggests Skills are potentially better than MCPs because they integrate tooling directly and reduce extraneous context loading.

![Screenshot at 00:01: The announcement slide from Anthropic introducing 'Equipping agents for the real world with Agent Skills,' setting the stage for the new capability pattern.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-00-01.png)

**Context:** The video discusses Anthropic's introduction of 'Agent Skills,' a novel mechanism designed to enhance the capabilities of Claude AI agents by providing them with structured, on-demand access to procedural knowledge, code, and resources relevant to specific tasks. This system aims to offer a more composable and token-efficient alternative to existing methods like Manifest Component Protocols (MCPs), which often require loading large amounts of context into the agent's prompt.

## Detailed Analysis

The presentation details Anthropic's new Agent Skills system, positioning it as a potentially superior alternative to MCPs for equipping agents with specialized knowledge. A Skill is fundamentally composed of a prompt, contained in a SKILL.md file, and optionally, tools and assets housed in a 'templates' subfolder, as illustrated by the 'algorithmic-art' skill structure (00:09). This structure dramatically improves token efficiency; for instance, one example showed context usage dropping from 4.2k tokens to just 70 tokens when using a Skill instead of loading all available MCP tools (1:40-1:55). The video demonstrates practical applications, including creating a 'slack-gif-creator' skill that utilizes pre-defined Python functions for creating GIFs adhering to Slack's constraints (2:52), and building a 'frontend' skill that codifies best practices for SuperDesign platform UI components, referencing internal documentation like 'component-guide.md' (4:19). The system allows agents to selectively load only the necessary files, such as specific tools or reference guides, based on the task at hand, leading to more complex and self-improving agent behavior, as demonstrated when creating a new frontend component (4:03). The speaker concludes by highlighting that this method allows developers to create skills based directly on their own codebase, fostering continuous improvement.

### Agent Skills Introduction

- Skills are introduced as a new pattern for specialized agents using organized folders of instructions, scripts, and resources
- A Skill equals Prompt (SKILL.md) + Tools & Assets (templates folder)
- Skills can be prompt-only, like the 'brand-guidelines' example.

### Token Efficiency Advantage

- MCP tools context usage of 4.2k tokens is reduced to 70 tokens when using a Skill structure for the 'shadcn' toolset
- This efficiency allows agents to handle more skills without exceeding context limits.

### Custom Skill Creation (GIF Creator)

- The 'slack-gif-creator' skill uses pre-defined Python functions in its 'core' folder (easing.py, color_palettes.py) to generate GIFs based on user prompts
- When prompted to create a GIF, the agent calls the skill, which executes the Python script to produce the final 'daily_standup_time.gif' (3:15).

### Custom Skill Creation (Frontend Best Practices)

- A 'frontend' skill is created within the 'ui-design' directory to enforce best practices for the SuperDesign Platform monorepo
- The skill references internal documentation files like 'component-guide.md' and 'styling-guide.md' to guide component creation (4:27).

### Self-Improving Agents

- By embedding project-specific knowledge and conventions (like the frontend component location strategy) directly into Skills, agents can continuously improve their adherence to internal standards without requiring the entire codebase in context.

![Screenshot at 00:01: The announcement slide from Anthropic introducing 'Equipping agents for the real world with Agent Skills,' setting the stage for the new capability pattern.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-00-01.png)
![Screenshot at 00:09: Diagram illustrating the composition of a Skill: SKILL.md serves as the Prompt, while the 'templates' folder contains Tools & Assets.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-00-09.png)
![Screenshot at 01:40: Token usage breakdown comparing System, Tools, MCP tools, and Custom Agents, showing the large token footprint of MCP tools.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-01-40.png)
![Screenshot at 02:21: The 'Shadon' MCP server settings showing that 7 tools are enabled by default, highlighting the potential for context overload.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-02-21.png)
![Screenshot at 03:37: Demonstration of the 'algorithmic-art' skill rendering generative art using p5.js based on the instructions within its Skill files.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-03-37.png)
![Screenshot at 04:04: The 'skill-creator' skill's documentation outlining best practices for creating new frontend components, including steps for package export and build processes.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-04-04.png)
![Screenshot at 04:44: The VS Code file explorer showing the nested structure of the custom 'frontend' skill, including references and styling guides.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-04-44.png)
![Screenshot at 05:01: The speaker points to the GitHub repository 'awesome-claude-skills,' encouraging community contributions via pull requests.](https://ss.rapidrecap.app/screens/1WImBwiA7RA/00-05-01.png)