All skills
Skillintermediate
Implementation
<div align="center">
Claude Code Knowledge Pack7/10/2026
Overview
<div align="center">
<div align="center">
š Home ⢠š§ Implementation
</div>Implementation
How Claude Code implements agentic patterns through components and architecture
Components vs Building Block
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
ā ā
ā š¦ BUILDING BLOCK (Anthropic) COMPONENTS (Claude Code) ā
ā āāāāāāāāāāāāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāāāāāāā ā
ā ā
ā Augmented LLM š¦ Subagent ā
ā (LLM + Retrieval + Tools + Memory) 𦓠Slash Command ā
ā š Skill ā
ā ā Foundation concept šŖ Hook ā
ā ā
ā Each Component internally uses ā Abstractions that organize ā
ā the Building Block to function agent capabilities ā
ā ā
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
Navigation
| Section | Description |
|---|---|
| components/ | š¦ Subagent, 𦓠Slash Command, š Skill, šŖ Hook |
| architecture/ | 5-Layer system architecture |
Quick Reference
Components Overview
| Component | Emoji | Invocation | Location |
|---|---|---|---|
| Subagent | š¦ | Task tool (šŖŗ) | .claude/agents/*.md |
| Slash Command | 𦓠| User /command | .claude/commands/*.md |
| Skill | š | Context-based | .claude/skills/*/SKILL.md |
| Hook | šŖ | Event-driven | .claude/settings.json |
Architecture Overview
| Layer | Emoji | Role |
|---|---|---|
| User Layer | šāāļø | Entry point |
| Main Agent Layer | š | Orchestration |
| Delegation Layer | š | Workflow definition |
| Execution Layer | ā” | Actual work |
| State Layer | š¾ | Persistence |
File Structure
.claude/
āāā agents/ # š¦ Subagent definitions
ā āāā *.md # One file per subagent type
āāā commands/ # 𦓠Slash Command definitions
ā āāā *.md # One file per command
āāā skills/ # š Skill definitions
ā āāā skill-name/ # One directory per skill
ā āāā SKILL.md # Skill content
āāā settings.json # šŖ Hooks and configuration
<div align="center">
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
</div>