Showing 20 of 20 concepts
- 01Configuration
CLAUDE.md
Project-level instructions Claude reads when beginning work.
7 MIN - 02Automation
Skills
Reusable instruction bundles loaded when a matching task requires them.
7 MIN - 03Agents
Subagents
Focused agents that handle independent tasks with separate context.
7 MIN - 04Tools
MCP
A standard that connects Claude Code to external tools and data.
7 MIN - 05Automation
Hooks
Automated commands triggered by Claude Code lifecycle events.
7 MIN - 06Workflow
Commands
Built-in and custom slash commands for repeatable actions.
6 MIN - 07Workflow
Plan Mode
A read-only workflow for analyzing work before changing files.
6 MIN - 08Context
Context Window
Claude's active working memory for the current session.
6 MIN - 09Context
Compaction
Summarizes older context to make room for continued work.
6 MIN - 10Security
Permissions
Rules that control which actions Claude can execute.
7 MIN - 11Workflow
Worktrees
Separate Git checkouts that support isolated parallel work.
6 MIN - 12Performance
Prompt Caching
Reuse of stable prompt prefixes to reduce latency and token cost.
7 MIN - 13Tools
Tools
Capabilities Claude uses to inspect and act on the environment.
7 MIN - 14Automation
Non-Interactive Mode
Running Claude Code in scripts, pipelines, or CI without a chat session.
6 MIN - 15Workflow
Checkpointing
Restore points that preserve useful states during development.
6 MIN - 16Configuration
Output Styles
Presets that change Claude's response format, tone, and role.
7 MINVERSION NOTE - 17Configuration
settings.json
Layered configuration for Claude Code behavior.
6 MIN - 18Performance
Background Bash
Long-running shell commands that continue while Claude works.
6 MIN - 19Context
Extended Thinking
More deliberate reasoning for difficult or ambiguous problems.
7 MIN - 20Security
Sandboxing
Filesystem and network restrictions around commands and tools.
7 MIN