Iterative Retrieval Pattern

The Iterative Retrieval skill addresses the 'context problem' in multi-agent AI workflows where sub-agents often lack the necessary project-specific knowledge to begin tasks effectively. By implementing a sophisticated four-phase cycle—Dispatch, Evaluate, Refine, and Loop—this skill allows Claude to start with broad queries and systematically narrow down to the most relevant files. This prevents context window overflow while ensuring the agent has access to critical implementation patterns, domain terminology, and hidden dependencies required for high-quality code generation.

主要功能

使用场景