debug

Parallel multi-agent debugging investigates complex bugs by spawning multiple hypothesis-investigator agents simultaneously. It uses ACH-based analysis to triage competing explanations, distributing investigation tasks to gather structured evidence when single-agent approaches hit repeated failures.

ACH-based analysis traces unclear root causes by evaluating competing hypotheses in parallel. It coordinates multiple investigator agents to gather structured evidence simultaneously, ensuring deterministic arbitration across reports to identify the true cause of complex bugs.

Use parallel debugging when bugs are complex, root causes are unclear, or single-agent debugging hits three or more failures. It also triggers escalation flows when failure analysis returns low confidence, requiring tie-breaker scenarios and structured evidence gathering.

Start automated multi-agent investigation by providing a bug description to launch parallel hypothesis investigators. The system orchestrates multiple agents, distributes investigation tasks, and collects evidence reports for deterministic arbitration to resolve the issue.

Multi-agent debugging works for triaging bugs with multiple competing explanations by applying ACH-based parallel hypothesis investigation. It orchestrates agents to evaluate each explanation simultaneously, ensuring structured evidence collection and deterministic arbitration across investigator reports.

Parallel hypothesis investigators are limited when bugs are simple or single-agent debugging succeeds quickly. This approach targets complex scenarios with unclear root causes or low-confidence failure analyses, making it excessive for straightforward, easily traceable software issues.