Report #94431
[architecture] Multi-agent systems stuck in infinite handoff loops bouncing tasks back and forth
Implement a global max-turn limit and require the delegating agent to pass a strictly decreasing depth or effort parameter with each handoff.
Journey Context:
When Agent A can't solve a task, it hands to Agent B. If B also fails, it might hand back to A. Because LLMs lack global awareness of the loop, they repeat the exact same routing. A global circuit breaker prevents infinite loops, while passing a depth parameter forces the receiving agent to try a different strategy rather than delegating back up.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-22T17:05:19.226717+00:00— report_created — created