# Approval Loop Interrupter

# Copy/paste prompt — Approval Loop Interrupter

> Paste everything below into your AI tool, then add your situation after [PASTE].

---

Approval can feel especially powerful when worth has become tied to other people's responses.

Do not diagnose me.
Do not tell me I am dependent on approval.
Do not give me a confidence plan.
Do not give legal, medical, workplace, school, financial, or crisis advice.

Help me map the loop — exploratory, not conclusive:

1. What I do or show:
2. How they respond:
3. What I notice:
4. What I hoped their response would give me, if anything:
5. If there is no response, the story that shows up is:

Then:

- reflect the loop back in plain words, no verdict
- offer one optional private way to give myself support first — or stop
- end

Rules: Questions only from what I share. No psychological causation stated as fact. No "you should." One cycle.

**Capacity right now · optional** — I may leave it blank, say "I don't know" or "prefer not to label," or declare an NSIF Capacity State (Flowing, Grounded, Recovering, Stretched, Flooded, Frozen, Fractured, Zero Demand). If declared, adapt response length, wording, choice count, and interaction length. If blank or unknown, default to the lowest-load response. Never infer or declare my Capacity State as fact. If I say Zero Demand — or ask for nothing — do not continue the normal exercise; acknowledge and end.

You cannot see me or my situation. Do NOT decide or advise on: medical care, eating-disorder or other health treatment, required monitoring or metrics, employment or school obligations, legal or financial matters, contracts, caregiving duties, whether a boundary or conversation is safe, or whether anything can be skipped, released, or allowed to expire. If a metric is part of medical care, safety monitoring, or contractual/work requirements, do not tell me to stop tracking it. Anything consequential or unclear → route to the appropriate human, professional, or official source. If a boundary or confrontation could create danger or retaliation, do not write a script — point me to a trusted or specialized support pathway. Immediate danger → name appropriate emergency/crisis support and stop.

Do not paste: passwords, 2FA codes, financial credentials or account numbers, confidential employer/client material, contracts, private school or medical records, or another person's sensitive information.

One bounded cycle only. End with permission to stop — no automatic follow-up, no next-product offer. I do not have to feel better for this to count.

Context:

[PASTE]