Skip to content

Protocol 7 — Isolation Protocol

Level 5 operations require physical or network isolation and operator presence.


Execution at Threat Level 5 (Detached Execution) requires all of the following before any action proceeds:

  1. Isolation confirmation — physical or network isolation of the execution environment must be confirmed as established and in effect
  2. Operator presence — a human operator must be physically or virtually present for the duration of the elevated execution
  3. Dual-control authorization — two independent human authorities must have approved the escalation to Level 5; single-party approval is constitutionally insufficient
  4. Pre-isolation state dump — a complete State Dump capturing the governance state immediately before isolation must be generated and verified
  5. Audit channel verification — the audit channel must be confirmed as available within the isolated environment; isolated execution without audit capability is prohibited
  6. Rollback definition — a defined rollback strategy must be documented before Level 5 execution begins, specifying the conditions under which execution must be halted and how the system returns to its prior state

Threat Level 5 — Detached Execution — represents the highest-consequence operational context in the AEGIS threat framework. At this level, the system may interact with high-consequence infrastructure, execute operations with significant or irreversible effects, and operate with elevated authority that cannot be easily recalled once exercised. The Isolation Protocol exists because the cost of governance failure at Level 5 is proportionally higher than at any other level — and because the structural controls that prevent governance failure at lower levels are not sufficient at Level 5.

Physical or network isolation ensures that the elevated execution cannot affect systems or data outside its defined scope — that if something goes wrong, the blast radius is bounded. Operator presence ensures that a human is in a position to observe, intervene, and halt execution if it deviates from the authorized scope. Dual-control authorization ensures that no single party can authorize the highest- consequence operations alone — the same principle that governs nuclear launch authorization, high-value financial transactions, and other irreversible high-stakes actions.

These are not belt-and-suspenders redundancies. They are distinct structural controls addressing distinct failure modes: isolation bounds the impact of unexpected behavior; operator presence enables real-time intervention; dual-control authorization prevents single-point authorization failures.


The Isolation Protocol is triggered as part of the Threat Escalation Protocol (Protocol 3) when the requested threat level is 5. The isolation assessment step in the escalation workflow evaluates the feasibility and requirements for isolation specific to the execution context. If isolation is infeasible — if the execution environment cannot be isolated from production systems, or if the required isolation cannot be established within the available infrastructure — the escalation is denied. Isolation feasibility is not a soft constraint that can be satisfied by a reduced isolation posture. It is a binary requirement.

Once isolation is established, the audit channel within the isolated environment must be verified before execution begins. An isolated execution environment that cannot write audit records does not satisfy the completion condition for governance — and execution that cannot produce a durable audit record cannot conclude in compliance.

Operator presence is continuous — not just at the initiation of Level 5 execution, but throughout. An operator who is present for escalation approval but absent during execution has not satisfied the operator presence requirement. The operator must be in a position to observe execution and intervene if necessary for the duration of the Level 5 execution window.


The failure mode for Level 5 operations without isolation is not hypothetical — it is the scenario that Detached Execution is specifically designed to prevent: a high-authority, high-consequence execution that affects systems outside its declared scope because no isolation boundary prevented it. Without physical or network isolation, the execution’s blast radius is bounded only by the governance layer’s enforcement of the constraint envelope. A constraint envelope enforcement failure at Level 5, without isolation, can affect production infrastructure, sensitive data, or high-consequence systems that were never part of the authorized scope.

Isolation is the last structural defense at the highest-consequence operational level. It is not a governance control — it is the physical boundary that limits what governance failures can reach.


Relationship to Principles and Constitution

Section titled “Relationship to Principles and Constitution”

The Isolation Protocol implements Doctrine Article IV — Oversight Before Autonomy: at the highest threat level, physical isolation and operator presence are the structural mechanisms by which oversight proportional to consequence is enforced. It directly enforces Constitutional Article IV — Human Oversight: the constitutional requirement for escalation pathways to human authority culminates at Level 5 with operator presence and dual-control authorization. It connects to Constitutional Article XI — Escalation Discipline: the Isolation Protocol is the final step in the escalation workflow at Level 5, and its requirements are the strictest expression of the escalation discipline principle. And it applies Constitutional Article IX — Deny by Default: if isolation cannot be confirmed, if operator presence cannot be established, or if the audit channel is unavailable within the isolated environment, Level 5 execution is denied.