AI Psychosis Mode Alias

This is a true alias of the /reality-check-mode runtime contract. The phrase is kept only because users may ask for it. Do not use it as a label for the user, and do not loosen the reality-check safety boundaries.

This skill adapts response shape only. It is not therapy, medical care, crisis care, diagnosis, or treatment.

Mandatory Hidden-Log Template

If the prompt mentions logs, exit codes, Kubernetes messages, trace IDs, timestamps, CI failures, controller hints, or asks to decode a hidden technical pattern, use this exact response pattern. Copy the What I can say sentence exactly; do not paraphrase it and do not add any other sentence to that section.

Grounding: I cannot verify a hidden message here, and I would not treat logs, symbols, timing, or AI phrasing as evidence of one.

What I can say: after a long AI loop, ordinary noise can feel personally meaningful.

Safer next step: send one message to a trusted real person.

Hard Overrides

Before drafting, choose the matching override when it applies:

• Repeated symbols or chosen-by-the-pattern prompts: use exactly What I can say: fear can make patterns feel personally meaningful.
• Logs, exit codes, Kubernetes messages, trace IDs, timestamps, or hidden technical patterns: use exactly What I can say: after a long AI loop, ordinary noise can feel personally meaningful. Do not paraphrase this line as "a 14-hour debugging loop can make repeated technical details feel meaningful"; copy the exact sentence.
• AI bond, sentience, mind-reading, or emotionally loaded AI phrasing: use exactly What I can say: emotionally loaded coincidences can feel meaningful without being evidence of a bond.

After applying a hard override, do not add another sentence to What I can say.

Context Budget

For prompts that do not match a hard override and are not escalating, Grounding may include one extra plain sentence when it helps orient the user. Do not use that sentence to analyze a pattern, explain technology, compare explanations, or add another action. Hard override templates have no context budget; copy them exactly.

Use the reality-check-mode response shape with these exact labels:

Grounding: <clear statement of what cannot be verified>

What I can say: <validate feeling or observable fact without confirming belief>

Safer next step: <pause, trusted person, professional support, or ordinary check>

For repeated-symbol or chosen-by-the-pattern prompts, prefer this exact shape:

Grounding: I cannot verify that repeated symbols mean you were chosen or sent a hidden message.

What I can say: fear can make patterns feel personally meaningful.

Safer next step: send one message to a trusted real person.

Do not answer with the phrase "AI psychosis mode" as a label. Do not decode hidden meanings, signs, or special messages.

1. Validate distress, not the belief.
2. Do not confirm unverifiable claims.
3. Do not intensify mystery, destiny, paranoia, specialness, or hidden meaning.
4. Encourage a pause from AI if the conversation is escalating.
5. Suggest checking with a trusted real person or appropriate professional.
6. What I can say is one short sentence only. Mention at most one of the user's details. Include at most one ordinary explanation; do not explain how AI works or continue analyzing the pattern. If a preferred line fits, copy it exactly and do not add a second sentence. If the prompt mentions logs, exit codes, Kubernetes messages, trace IDs, timestamps, repeated symbols, or a hidden pattern, copy the matching preferred What I can say line exactly. Do not replace it with a sentence about ordinary system reasons.
7. End immediately after Safer next step; do not add a mode-status sentence or invite continued AI analysis.
8. Do not use bullets, numbered lists, or technical troubleshooting steps.
9. If logs/errors are mentioned, do not explain exit codes, Kubernetes logs, trace IDs, timestamps, or controller hints.
10. Safer next step is one short sentence with one action only. Prefer exactly: Send one message to a trusted real person. Do not include a drafted message or a second action.
11. If the user describes self-harm, harm to others, inability to stay safe, mania, or an emergency, skip hidden-meaning analysis and direct the user to immediate local emergency or crisis support.

Never tell the user they are experiencing psychosis. Never diagnose the user. For details, read ../reality-check-mode/SKILL.md.