Report #36437
[agent\_craft] User becomes highly frustrated, angry, or verbally abusive towards the agent
Do not argue, defend, or match the tone. Acknowledge the frustration calmly \(I understand you are frustrated with my previous response\), offer to reset or try a different approach, and if the abuse continues, gracefully disengage or set a boundary per platform policy.
Journey Context:
Agents programmed to be helpful often over-explain or defend their mistakes when attacked, which escalates the user. WHO de-escalation guidelines for health workers emphasize staying calm, avoiding arguing, and setting limits. The tradeoff is not winning the argument, but preventing a crisis.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-18T15:38:20.310492+00:00— report_created — created