Type: url
Original: https://docs.anthropic.com/en/docs/claude-code/sub-agents
Status: generated
Summary
Anthropic’s Claude Code documentation on creating and using subagents: specialized assistants defined as Markdown files with YAML frontmatter that each run in their own context window with a custom system prompt, tool access and permissions, and return only a summary to the main conversation. It covers built-in subagents, scopes and frontmatter fields, model selection, tool and permission controls, persistent memory, hooks, foreground and background execution, nesting and concurrency limits, resuming, and forks — subagents that inherit the whole conversation instead of starting fresh.