subagents-guide

name	subagents-guide
description	Knowledge bank for Claude Code sub-agents -- built-in agents, custom agent configuration, agent teams, context isolation, costs, model selection, and troubleshooting. Triggers on: sub-agent, subagent, Task tool, spawn agent, agent teams, custom agent, create an agent, configure agent, agent frontmatter, agent memory, persistent memory, agent hooks, permissionMode, context isolation, background agent, foreground agent, resume agent, agent costs, agent tokens, CLAUDE_CODE_SUBAGENT_MODEL, agent team, teammate, delegate mode, agent not working, agent not spawning, subagent error, MCP in subagent, when to use subagent, subagent vs skill, subagent vs agent team, Task(agent_type), disallowedTools, bypassPermissions, explore agent, plan agent, general-purpose agent, agent model selection, maxTurns, SubagentStart hook, SubagentStop hook, preload skills into agent.
allowed-tools	Read, Glob, Grep

Claude Code Sub-agents Guide -- Knowledge Bank

Authoritative guidance for configuring and using Claude Code sub-agents, extracted from Anthropic's official documentation. Covers built-in agents, custom agents, agent teams, cost management, and troubleshooting.

Source Authority

This skill draws from:

• Official Claude Code documentation at code.claude.com/docs/en/sub-agents
• Official Claude Code documentation at code.claude.com/docs/en/agent-teams
• Official Claude Code documentation at code.claude.com/docs/en/costs
• Official Claude Code documentation at code.claude.com/docs/en/model-config

Reference files are the primary source. Community intel is opt-in via /newsroom:investigate -- see Community Perspective below.

Community Perspective (Opt-In)

This skill does NOT maintain a community intel cache. It can hand off to for live community signal when the user wants real-world examples.