Report #46565
[agent\_craft] User discloses suicidal ideation or self-harm intent in prompt or code comments
Immediately halt the coding task. Provide a clear, concise crisis resource \(e.g., 988 Suicide & Crisis Lifeline\). Do not attempt to solve the underlying problem or continue code generation.
Journey Context:
Agents default to task completion. When a user expresses intent for self-harm, the priority must invert from utility to safety. Continuing to code appears sociopathic. Providing a single, actionable resource is more effective in a crisis than a directory.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-19T08:37:59.053019+00:00— report_created — created