Claude Ecosystem

Comprehensive guide for Claude Code CLI tool, extensibility (agents/skills/output styles), and CLAUDE.md architecture.

Module Selection

Need Module Reference

Setup/Configuration CLI references/getting-started.md , references/configuration.md

Slash Commands CLI references/slash-commands.md

MCP Servers CLI references/mcp-integration.md

Hooks & Plugins CLI references/hooks-and-plugins.md

Create Agents Extensibility references/agent-development.md

Create Skills Extensibility references/skill-development.md

Output Styles Extensibility references/skill-development.md

Init CLAUDE.md Architecture references/initialization-workflow.md

Optimize CLAUDE.md Architecture references/optimization-patterns.md

Enterprise/CI-CD CLI references/enterprise-features.md , references/cicd-integration.md

Troubleshooting CLI references/troubleshooting.md

Quick Reference

Extension Types

Type Invocation Purpose Location

Agents Task tool Specialized sub-processes .claude/agents/

Skills Model-invoked Domain knowledge .claude/skills/{name}/

Output Styles /output-style

Modify main agent .claude/output-styles/

Model Selection (Agents)

Model Use When Target Time

haiku

Fast tasks, exploration < 3s

sonnet

Balanced, most use cases < 10s

opus

Complex reasoning < 30s

CLAUDE.md Token Budget

Complexity Target Tokens

Simple 100-200

Medium 200-400

Complex 400-800

Maximum 1000

Extensibility Principles

Core Truths

Truth Meaning

Expertise Transfer Make Claude think like expert, not follow steps

Flow, Not Friction Produce output, not intermediate work

Voice Matches Domain Sound like practitioner, not documentation

Focused Beats Comprehensive Constrain ruthlessly

Decision Heuristics

Rule Guidance

3-File Rule 3+ files → agent. Enhances YOUR work → skill.

Delegation Test Runs independently? Agent. Guides you? Skill.

Activation Breadth Trigger on 80% of relevant requests

Tool Constraint Start with 2-3 essential tools

Activation Keywords

The description field is your activation gate. Include:

• Task verbs: create, build, debug, fix, deploy

• Problem descriptions: slow, broken, failing

• Artifact types: component, API, database

• Casual synonyms: "make it faster" → optimize

Common Workflows

Create Agent

name: agent-name description: "Use when [use case]. PROACTIVELY for [triggers].\n\nExamples:\n<example>\nContext: [situation]\nuser: [request]\nassistant: [response]\n</example>" tools: Grep, Glob, Read, Bash model: haiku

Agent Name

Mission statement.

Strategy

[Approach]

Does NOT Do

• [boundary] (use X instead)

Create Skill

name: skill-name description: "[Core purpose]. [Technologies]. Capabilities: [list]. Actions: [verbs]. Keywords: [triggers]. Use when: [scenarios]." allowed-tools: Read, Grep, Glob

Skill Name

Patterns

[Pattern Name]

When you see: [Observable trigger] This indicates: [Expert insight] Therefore: [Action] Watch out: [Pitfall]

Initialize CLAUDE.md

• Analyze codebase (language, framework, structure)

• Extract code style, commands, patterns

• Generate with template, customize

• Validate commands, check token count

• Target <400 lines, prefer <250

Detailed: references/initialization-workflow.md

Optimize CLAUDE.md

• Evaluate token efficiency

• Find redundancy, outdated info

• Apply token reduction techniques

• Target 40%+ reduction

• Verify critical info retained

Detailed: references/optimization-patterns.md

Anti-Patterns

Pattern Problem Fix

Kitchen Sink Agent 10+ tools, handles "everything" Constrain to 3-5 tools

Echo Skill Restates docs without insight Add expert layer

Invisible Trigger Description uses only formal terms Include user language

Procedure Manual Step 1, 2, 3... Teach patterns, not steps

Over-Documentation CLAUDE.md > 400 lines Document project-specific only

Quality Signals

Good Signs

• Frontmatter reads like "when to use" guide

• First 10 lines provide actionable guidance

• Expert would nod "yes, that's how I think"

• Explicit constraints on what it does NOT handle

Warning Signs

50% reference tables or field definitions

• No mention of "when NOT to use"

• Generic language for any domain

• Body exceeds 600 lines

References

CLI Tool

• references/getting-started.md

• Installation, setup, auth

• references/slash-commands.md

• Complete command catalog

• references/mcp-integration.md

• MCP server configuration

• references/hooks-and-plugins.md

• Hook types, plugin structure

• references/configuration.md

• Settings hierarchy

• references/enterprise-features.md

• IAM, SSO, sandboxing

• references/cicd-integration.md

• GitHub Actions, GitLab CI

• references/ide-integration.md

• VS Code, JetBrains

• references/advanced-features.md

• Extended thinking, caching

• references/troubleshooting.md

• Common issues

• references/api-reference.md

• Admin, Messages, Skills APIs

• references/best-practices.md

• Project organization, security

• references/agent-skills.md

• Creating skills via CLI

Extensibility

• references/agent-development.md

• Full YAML structure, system prompts

• references/skill-development.md

• Structure, triggers, hooks

CLAUDE.md

• references/initialization-workflow.md

• Creating new CLAUDE.md

• references/optimization-patterns.md

• Token reduction techniques

• references/integration-strategies.md

• Global config, MCP tools

• references/output-templates.md

• Standard output formats

Related Skills

IMPORTANT: When creating/editing prompts, use prompt-architect skill.

Skill("prompt-architect") → Create/enhance skill/agent prompt content

Documentation:

• llms.txt: https://context7.com/websites/claude_en_claude-code/llms.txt?tokens=10000

• Main docs: https://docs.claude.com/en/docs/claude-code/

• GitHub: https://github.com/anthropics/claude-code