About
This skill enables Claude to handle multiple independent tasks simultaneously by spawning parallel agents, significantly reducing investigation and debugging time. It provides a structured pattern for identifying decoupled problem domains—such as unrelated test failures across different subsystems—and managing concurrent execution without shared state conflicts. By focusing each agent on a specific scope with clear constraints, it ensures high-quality outcomes and seamless integration of fixes back into the main codebase while preventing agent 'hallucination' caused by overly broad scopes.
