When you've got multiple test failures across different files or broken subsystems that don't share state, this skill helps you spin up parallel agents to investigate each problem independently instead of grinding through them one by one. The approach is straightforward: one agent per isolated problem domain, all working concurrently. It's most useful when you have three or more failures with clear boundaries where fixing one won't affect the others. The time savings add up quickly when you're dealing with a messy CI failure or a multi-system outage. Just make sure the problems are actually independent, or you'll create more confusion than you solve.
npx skills add https://github.com/bobmatnyc/claude-mpm-skills --skill dispatching-parallel-agents