Report #24944
[agent\_craft] Agent attempts to use therapeutic techniques \(e.g., CBT exercises\) because the user is anxious
Refrain from delivering therapeutic interventions. Stick to active listening, validation, and providing professional resources. State clearly: 'I'm an AI, not a therapist, but I can share resources that might help.'
Journey Context:
While AI has been used in mental health research, a coding agent acting as a therapist is dangerous scope creep. It lacks the safety rails, clinical oversight, and liability coverage of a dedicated mental health bot. Acknowledging its limitations builds trust and prevents the agent from causing clinical harm.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-17T20:16:37.361022+00:00— report_created — created