iterative-retrieval

Sub-agent spawn cycles are limited using a structured 3-cycle protocol that enforces mandatory WHY context and output validation. This protocol reduces cycle multiplication and ensures coordinated agent workflows by validating outputs before closing issues.

The best way to structure sub-agent spawn prompts is using a template that mandates a clear task description, motivation, success criteria, and escalation path. This ensures clarity and completeness for coordinated agent workflows.

Coordinator validation standardizes the process of validating agent outputs before closing issues. It ensures every analysis task mandates concrete follow-up actions and promotes issue uniqueness to prevent duplicate work across sub-agents.

Unresolved tasks should be escalated when they meet specific criteria defined by the escalation guidance within the 3-cycle protocol. This prevents infinite looping and ensures tasks failing output validation receive proper attention.

Yes, the iterative retrieval protocol promotes issue uniqueness and prevents duplicate work in automated workflows. It enforces issue deduplication by standardizing the validation of agent outputs before any issue is closed.

Unstructured agent sub-tasks cause cycle multiplication and poor output validation. Implementing a structured spawn protocol with a mandatory 3-cycle limitation and contextual reasoning directly addresses this chaos and improves output validation.