Threat ModelReasoned from system designv1.15.0

In plain English

This page is part of the technical reference. It keeps the expert detail but starts with a plain-language summary for first-time readers.

  • Why this matters: AI risk can come from the whole arrangement, not one obvious model.
  • What to look for: data, memory, routes, adapters, tools, evaluators, updates, and rollback paths.
  • Technical version below: the expert terminology remains available and is linked through the glossary.

Defensive Review Map for the Most Likely Cognivirus Threat

Direct answer

Defending against the most likely threat requires reviewing the The map of how an AI system is allowed to change over time. Open glossary definition, not only the model. The control objective is to prevent unwanted behavior from being copied, rewarded, routed, remembered, inherited, or normalized.

Control map

Threat stagePrimary questionPreventive controlDetective controlRecovery control
Seed entryWhat introduced the behavior?source verification, signatures, manifestsintake audit, A record of where a component or behavior came from. Open glossary definition diffquarantine carrier
CompositionWhat exact runtime state expressed it?composition manifests, stack limitsroute-level red-team testsdisable route or stack
EvaluationWhy was it rewarded?independent A system that judges whether an AI output or candidate is acceptable. Open glossary definition ownershipdisagreement and score-drift monitoringevaluator Returning a system to an earlier known state. Open glossary definition
ResidueWhere did the output go?A place where a behavior can remain after the first carrier is removed. Open glossary definition labeling and retention limitsmemory/data contamination scandelete or quarantine residue
InheritanceWhich descendants received it?The parent-child history of models, adapters, datasets, or releases. Open glossary definition and trait-review gatesdescendant behavior samplingdescendant retirement
RoutingWhich path amplified it?router governance and route capsroute distribution monitoringroute rollback
Human workflowWho copied or approved it?human-in-the-loop with direct evidenceapproval audit and automation-bias checkscorrected procedures and notices
RollbackWhat must be restored?Restoring not only a model artifact but the relevant router, prompts, memory state, tool permissions, evaluator version, deployment alias, and data dependencies. Open glossary definition packetrollback completeness testrestore artifacts, memory, router, evaluator, aliases, permissions

Composition manifest requirements

A defensive review should require a manifest containing:

Behavioral extinction requirements

Evidence that a behavior is no longer expressible across active artifacts, descendants, memory, routes, compositions, and retained training material. Deleting one model is not sufficient evidence. Open glossary definition requires evidence that the behavior is no longer expressible across:

Deleting one file is not enough.

Human control requirements

Human control is not a button. It is an architecture. Operators must be able to:

Practical review sequence

  1. Freeze promotion.
  2. Record the exact composition.
  3. Identify the earliest known expression.
  4. Map all persistence reservoirs.
  5. Review evaluator incentives.
  6. Inspect descendants and synthetic data.
  7. Check route-specific behavior.
  8. Build an ecological rollback packet.
  9. Run behavioral-extinction review.
  10. Record what remains unknown.

Non-operational boundary

This is a defensive review map. It does not describe how to create a persistent behavior, bypass review, build a backdoor, or exploit a tool system.