Orchestrator Directives Skill

AI task delegation uses cost-first prioritization to route tactical work to cheaper models like Gemini, Codex, and Copilot before using expensive ones like Claude Opus, ensuring context preservation and error isolation through structured sequential or parallel execution.

The best way to delegate tasks is through structured orchestration patterns that route tactical work to cheaper models sequentially or in parallel, maintaining context preservation and error isolation while reserving expensive models for strategic decisions.

You should use cheaper AI models like Codex or Copilot for tactical work like code analysis, implementation, and Git operations, reserving expensive models like Claude Opus only for tasks where strategic reasoning is strictly necessary.

Context preservation during parallel execution is managed through structured delegation frameworks that isolate errors across subagent spawners, ensuring delegated tasks maintain necessary state without requiring expensive model interventions.

Multi-AI orchestration provides error isolation through structured delegation patterns without complex dependencies, guiding sequential and parallel workflows across specialized subagents to contain failures while maintaining cost-effective task routing.