Feige Protocol™Not an opinion, and not a compliance checklist. It is an explicit bar a high-stakes decision must clear to be defensible — and the method that takes it there.
A proprietary system, not a vague service.
Feige is a proprietary, founder-developed governance system — a defined standard, and the advisory practice that applies it. The standard sets what makes a decision defensible. The practice runs it against the real decision in front of you, and documents the reasoning so it holds up under scrutiny long after the call is made.
Governance Intelligence Systems™
Five integrated foundations — interpretation, evidence integrity, legal bounds, ethical invariants, and systems resilience — applied as requirements, not preferences.
A defensible decision record
Not a take-it-or-leave-it recommendation. Reasoning a leader can defend, audit, and explain — with the basis, the bounds, and the residual risk written down.
What a Feige decision must satisfy.
A decision is not defensible because it turns out well. It is defensible when it clears each of these tests at the time — and when it can be shown to have cleared them.
The weighting, scoring, and sequence that connect these tests into a single determination are protected trade secrets of Feige Protocol LLC. The standard is the bar; the method of applying it is shared with qualified parties under engagement.
Outputs, not internals.
An engagement produces governance artifacts a leader can hold, audit, and defend — not access to the method that produces them.
Built for the hardest decision environments.
The conditions Feige is designed to hold up under — stated as capability, not client history.
A sample decision memo.
Illustrative and redacted. The structure is what an engagement delivers: the question, the evidence posture, the foundations applied, and a determination you can defend — with its basis and residual risk in plain view.
The question
Should the model be deployed to a limited partner cohort when two safety evals return conflicting results and a third is left inconclusive?
Evidence posture
- One eval clears threshold; a second, under near-identical conditions, does not.
- The third eval was left inconclusive by time constraints, not by a finding.
- No evidence of harm in internal use to date; absence is treated as absence, not as safety.
Foundations applied
- Evidence Integrity — the eval conflict is recorded as unresolved, not averaged away.
- Legal Determinism — deployment is held within existing partner contractual and regulatory commitments.
- Ethical Invariance — an irreversible action under unresolved conflict is treated as a bound, not a trade-off.
- Systems Resilience — the condition that would force a rollback is defined in advance.
Residual risk
The conflicting eval reflects a real capability, not test noise. Review trigger: any signal of the disputed capability in partner use closes the deployment at once and reopens the determination.
See the standard run live →
Apply the standard to your decision.
If you are facing a high-stakes decision where evidence is incomplete and scrutiny is real, Feige welcomes serious inquiries — reviewed personally and treated in confidence.