Constraint Registry

Not every experiment is allowed to proceed.

The Constraint Registry records formal limits imposed on experiments when risk, instability, or misalignment exceeds acceptable thresholds. Constraints are not punishments. They are protective measures.

Each registry entry includes:

  • The experiment or domain affected

  • The constraint imposed

  • The condition that triggered it

  • The authority or process that applied it

  • Whether the constraint is temporary or permanent

Constraints are timestamped and versioned. No constraint is applied silently.

Stopped Experiments

Some experiments are halted before completion.

Stopped experiments differ from failed experiments. Failure occurs when a system collapses under testing. Stoppage occurs when continuation would cause unacceptable harm, distortion, or misinterpretation.

Each stopped experiment record includes:

  • Original hypothesis

  • Progress at time of stoppage

  • Reason for intervention

  • What risk was identified

  • Whether the experiment may resume under new conditions

Stopping an experiment is treated as an outcome, not an omission.

Risk Thresholds

ZinxLabs defines explicit thresholds beyond which experimentation must pause or terminate.

Risk thresholds may include:

  • Misinterpretation risk

  • Cultural destabilization

  • Governance ambiguity

  • AI misuse potential

  • Temporal or continuity disruption

Thresholds are reviewed periodically and updated based on accumulated evidence. They are not fixed doctrines.

Refusal Notes

ZinxLabs documents refusals to pursue certain lines of inquiry.

Refusal notes exist to make absence visible. They record:

  • What was proposed

  • Why it was declined

  • What assumptions were rejected

  • Whether the refusal is permanent or conditional

Refusals prevent future repetition of unsafe or unproductive paths.

Constraint Propagation

Constraints applied in one area may affect others.

Constraint propagation pages track:

  • Which experiments or domains were indirectly limited

  • What shared assumptions triggered the spread

  • How protocols were updated across the lab

  • Whether new safeguards were introduced

Constraints are traced, not isolated.

Governance Interface

ZinxLabs does not govern itself in isolation.

This page documents how constraints interface with:

  • ZynxLogic interpretations

  • ZynxSecs governance protocols

  • ZynxSecurity enforcement boundaries

The purpose is not to externalize responsibility, but to show how experimental limits become institutional safeguards.