Parallel Agents

This orchestrates multiple specialized agents within Claude Code to tackle complex tasks that need different areas of expertise. You invoke agents sequentially or in parallel (security auditor, backend specialist, test engineer, etc.), pass context between them, and synthesize their findings into a unified report. The skill shines when you need comprehensive analysis from multiple angles, like reviewing a feature's security posture, API design, and test coverage all at once. It's overkill for simple changes but legitimately useful when you'd otherwise be context-switching between different review perspectives yourself. Seventeen domain-specific agents are available, and everything stays in one session with shared context.