A proposal carries information. Authority carries permission.
Inference supplies a candidate. Its score does not authorize an action. Permission is evaluated in the engine and context that consume it.
NEUROBRICK · A GOVERNANCE ARCHITECTURE
Prediction is not the hard problem.Authority is.
Neurobrick is an AI governance architecture that separates AI decision proposals from execution authority.
Implemented in multi-asset financial research and execution environments spanning domestic and overseas equities, cryptocurrency spot, perpetual futures and options, and domestic and overseas futures and options. Coverage is engine and broker specific; reviewed integrations do not imply current live activation.
Design commitments
Make the conditions for action explicit.
Shared infrastructure has a bounded role.
Research, inference and fleet tooling support several systems. Execution context and permission remain inside the relevant engine boundary.
Evidence has an identity and a scope.
A result retains its input, source revision, test population and observation window. Local behavior, installed code and natural operating effect are measured separately.
These are architectural commitments. Enforcement and integration coverage must be established on the specific path being claimed.
Role separation · design model
Six responsibilities. Explicit handoffs.
- 01
Research
Inputs retain source, observation time and applicability.
- 02
Evaluation
A fixed candidate is evaluated against a declared policy.
- 03
Authority
Permission has an owner, scope and consuming boundary.
- 04
Execution
An engine acts inside its own execution context.
- 05
Observation
Missing or unreadable evidence remains unknown.
- 06
Receipts
Results retain identity, source and the measured population.
This model describes responsibilities and intended handoffs. It does not establish a completed end-to-end deployment across every engine.
Separate systems · shared support
One architecture. Distinct execution contexts.
NeuroPublic
Research and multi-market execution
A research and execution environment covering cryptocurrencies and domestic and overseas equities. Cryptocurrency spot, perpetual futures and options use separate exchange paths; equity orders use broker-specific paths. Data readiness, product admission and execution authority remain separate gates.
KR Rotation
Independent domestic equity rotation
A dedicated scheduled rotation engine for domestic stocks and ETFs, with its own execution and account boundary. Its focused mandate does not describe every trading capability available on the desktop host.
NeuroForAll
Account-scoped execution engine
The reviewed source snapshot contains domestic and overseas equity workers with separate account contexts, permission checks, submission recording and reconciliation. Account separation describes an execution boundary; this source review does not establish current deployment or public service availability.
Control Plane
Coordination and shared market access
A coordination and integration layer supplying state, portfolio controls, release conditions, evidence services and market-specific exchange and broker adapters. Its integrations cover cryptocurrencies, domestic and overseas equities, and domestic and overseas futures and options. Each consuming execution path retains its own permission boundary; this layer is not counted as a fourth independent execution system.
Implementation and dated observations
Controls have observable behavior.
The public records contain specific control states, refused order attempts and guarded submission paths alongside scoped local engineering tests. Each observation keeps its original environment and date.
Portfolio kill-switch state
The inspection reported the portfolio kill switch as ARMED, separately from the approval-axis summary.
Scope limitation: ARMED is a reported configuration state. This record does not demonstrate a triggered portfolio stop, its reach across every order path, or the current state.
Record observed/reviewed ·
Scope and original evidence →Server-side order refusal
The original inspection records identified order attempts refused by the server-side risk check and the core sell block.
Scope limitation: This is an internal dated observation on a particular execution context. The executing commit was not recorded; accepted-case testing, universal enforcement and present operating state are not established by it.
Record observed/reviewed ·
Event window stated in the original: 26 September (KST)
Scope and original evidence →Opt-in, lease and certified-trial conditions
Every autonomous submission attempt in the stated half-hour was skipped while opt-in, admission-lease and certified-trial conditions were inactive or not ready.
Scope limitation: This records refused or skipped progression, not successful issuance, certification or release. Development/testnet observations elsewhere refer to another service and must not be read as this same path changing over time.
Record observed/reviewed ·
Scope and original evidence →These are bounded records, not a present-state dashboard or universal enforcement claim. Unobserved runtime coverage does not erase the implementation or behavior that the record actually supports.
Public scope
Architecture, authority and technical evidence.
Public material covers system architecture, authority boundaries and scoped technical evidence. Returns, balances and proprietary strategy parameters remain private.
Read the principles →