Report #71032
[architecture] Infinite loops of handoffs where Agent A delegates to Agent B, who delegates back to Agent A because neither can resolve the issue
Implement a global turn or handoff counter and a fallback supervisor agent that catches cycles and forces termination or user-escalation.
Journey Context:
LLMs lack global awareness of the conversation graph. They just see the last message, think 'not my job,' and pass it back. Without a circuit breaker, the system runs indefinitely, burning tokens. The tradeoff is that hard limits might terminate valid, complex reasoning chains, so the limit must be tuned to the specific domain, and the fallback agent must be capable of graceful degradation rather than just crashing.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-21T01:48:31.206801+00:00— report_created — created