dag-iteration-detector

To decide when to re-run a task in a DAG pipeline, analyze quality signals, validation results, confidence scores, and explicit feedback to identify if outputs require re-execution, then apply retry, refine, or escalate strategies.

Task iteration in DAG workflows is triggered by validation failures, low confidence scores, missing requirements, or explicit feedback, which signal that the current output quality is insufficient and needs rework.

You can manage iteration budgets for DAG task retries by configuring bounded execution limits that automatically select between retry, refinement, expansion, or escalation strategies to prevent unbounded resource consumption.

Yes, iteration detection integrates with dag-dynamic-replanner and dag-convergence-monitor to close the feedback loop, enabling coordinated re-execution decisions and pipeline convergence across your DAG-based task pipelines.

Detecting re-execution needs requires structured inputs including the completed task output, quality signals, and configuration parameters, which are processed to produce a structured iteration decision with triggers, strategy, and next steps.

The best way to handle failed validation in DAG task pipelines is to run an iteration detector that evaluates the quality signals and validation results to output a structured decision recommending whether to retry, refine, or escalate the task.