Agent Orchestration

关于

This skill provides a rigorous framework for Claude when acting as an orchestrator, enabling it to delegate complex tasks to specialist sub-agents with high precision. By leveraging the scientific method—focusing on raw observations, measurable success criteria, and 'world-building' context—it prevents the common pitfall of over-prescribing implementation details. This approach allows sub-agents to fully utilize their 200k context windows and specialized toolsets to discover optimal solutions, verify findings against authoritative sources, and maintain high standards across multi-agent development cycles.

主要功能

• Pre-delegation checklists to ensure context integrity and prevent rot
• 11 GitHub stars
• Standardized sub-agent delegation templates for role consistency
• Context calibration for both focused fixes and broad investigations
• Scientific method alignment for evidence-based task execution
• World-building resource patterns for describing complex environments

使用场景

• Managing multi-stage CI/CD and linting workflows with autonomous agents
• Coordinating complex refactors across multiple specialized sub-agents
• Delegating deep root-cause investigations of intermittent system bugs