The Agent Skills Directory

Agent Orchestration Pattern: The skill defines a procedural workflow for breaking down large tasks into smaller, manageable subtasks delegated to subagents. This is a standard functional pattern for enhancing the performance of AI agents on complex projects.
Feedback and Iteration Cycle: The instructions include requirements for the agent to review subagent plans, provide feedback, and iterate on produced code. This creates a human-in-the-loop or agent-in-the-loop validation process that helps ensure the quality and safety of the generated output.
Validation through Testing: The skill mandates the creation of tests for each subagent's task and high-level tests for the overall behavior. This technical requirement ensures that the subagents' work is verified against expected behaviors before being integrated into the final output.
Instructional Context Handling: The skill focuses on context management and delegation using platform-provided tools (such as the Task tool). It does not introduce external dependencies or perform network operations outside of its intended orchestrator role.

handle-large-tasks