Debate security audit findings with three competing subagents (Defender + Red Team + Compliance) across two rounds. Use when user says "debate security", "security from 3 perspectives", "is this vulnerability real", "security team review", "red team this", "challenge security findings", "false positive check". Best for 10+ findings where severity needs validation. Each agent runs in isolated worktree.
More Claude commands
/agents
Manage subagents
Create or manage subagents, or edit the agents directory directly.
/team-onboarding
Team onboarding
Generate a team onboarding guide from your usage history.
/deep-research
Deep research
Fan out web searches, cross-check sources, and synthesize a cited report.