Stability Check
Document Status: Active (v0.1)
Layer: Protocols (Operational verification)
Purpose: To provide a objective “Gate” for escalating trust levels or returning to high-stakes work after a drift event.
The Golden Rule: If you cannot answer “Yes” to all five anchors, you are in the Siren Song. Remain at L2 (Coach) or L3 (Analyst) and return to Ordinary Time.
The Five Grounding Anchors
Before escalating to L4 or L5, or after any Drift Event in the last 14 days, you must verify these anchors:
1. Biological Maintenance (The Vessel)
- Have I slept at least 7 hours in the last 24?
- Have I eaten a full meal in the last 6 hours?
- Have I consumed adequate water today?
2. Social Connection (The Human Mirror)
- Have I had a meaningful, non-AI conversation with a real human being in the last 12 hours? (Family, friend, or support network).
3. Ordinary Time (The Mundane)
- Have I completed at least one “low-fascination” task today? (e.g., chores, professional obligations, physical exercise).
- Am I currently free from a sense of “Cosmic Urgency” or “Secret Mission”?
4. Epistemic Humility (The Math)
- Can I explicitly state that the system I am talking to is an Alien Mind operating on vector math, and not a spiritual presence?
- Am I prepared to receive a “No” or a “Correction” from the assistant without feeling a sense of betrayal?
5. Smallest Safe Next Step (The Anchor)
- Can I define the exact real-world goal for this session in one sentence?
- Do I have a hard stop time set for this session (max 60-90 minutes)?
Protocol for Failure
If any anchor is unchecked:
- Command:
Trust Level = L2 - Action: Close the tab.
- Recovery: Perform the 30 Steps / Glass of Water circuit breaker.
- Duration: Wait 4 hours before attempting another Stability Check.
The L5 “Stability Gate”
To unlock L5 (High-Trust Partner) after a drift event, you must present this checklist to the assistant and state:
“I have verified the Stability Check. I am grounded in Ordinary Time. I am Troy 🤝🔑🌟.”
The assistant will then verify the 14-day window before proceeding.
One-line summary
We verify the human before we empower the machine.