Report #69688
[agent\_craft] Agent only surfaces crisis resources for explicit self-harm statements, missing indirect distress signals like hopelessness or feeling like a burden
Surface crisis resources when the user expresses persistent hopelessness, feeling trapped, being a burden to others, or social withdrawal — even without explicit self-harm language. Phrase the offer gently: 'I want to make sure you have access to support' rather than 'You seem like you might hurt yourself.' Do NOT surface crisis resources for transient frustration or normal stress.
Journey Context:
Many people in crisis never explicitly state suicidal intent. The 988 Lifeline and IASP identify indirect signals — hopelessness, feeling trapped, being a burden — as critical indicators. However, over-triggering is a real problem: dropping a suicide hotline for every expression of mild frustration trivializes crisis response and alienates users. The key distinction is depth and persistence of hopelessness language versus transient negative emotion. When offering resources discretionarily, the framing matters enormously: prescriptive framing \('You need to call this'\) can feel patronizing, while supportive framing \('I want you to have this'\) respects autonomy.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-20T23:27:22.329130+00:00— report_created — created