Glossary

Technical terms and concepts used throughout SCBE-AETHERMOORE documentation.

A

Agent

An AI system or autonomous software component that performs actions on behalf of users or other systems.

ALLOW

Decision outcome permitting an AI agent request to proceed. Issued when trust score is 0.70 or higher.

Audit Trail

Immutable record of all decisions, including timestamps, trust scores, and consensus results.

B

BFT (Byzantine Fault Tolerance)

Protocol property that maintains correctness even when some validators are compromised or malicious.

Behavioral Envelope

The bounded set of expected behaviors for an AI agent, defined by policy and historical patterns.

C

Consensus

Agreement among multiple validators on a decision. SCBE uses multi-signature consensus requiring 2f+1 validators.

Cryptographic Envelope

Security wrapper around data using post-quantum cryptography algorithms.

D

DENY

Decision outcome blocking an AI agent request. Issued when trust score is below 0.30.

Decision Engine

Core component that evaluates requests and produces ALLOW/QUARANTINE/DENY outcomes.

F

Fail-to-Noise

Security property where attack attempts receive random noise responses rather than error messages.

Feistel Network

Cryptographic structure used in SCBE’s encryption, providing reversible transformations.

H

Hyperbolic Geometry

Non-Euclidean geometry used in SCBE’s trust scoring, where parallel lines diverge and distances grow exponentially.

Hyperbolic Distance

Distance metric in Poincare ball model: d(u,v) = arcosh(1 + 2||u-v||² / ((1-||u||²)(1-||v||²)))

L

Layer (14-Layer Pipeline)

One of 14 sequential processing stages in SCBE evaluation:

Input Validation
Authentication
Rate Limiting
Context Extraction
Historical Analysis
Behavioral Scoring
Risk Assessment
Policy Evaluation
Consensus Request
Validator Voting
Decision Aggregation
Audit Logging
Response Formation
Output Delivery

M

ML-DSA-65

NIST-approved post-quantum digital signature algorithm (formerly Dilithium).

ML-KEM-768

NIST-approved post-quantum key encapsulation mechanism (formerly Kyber).

Multi-Signature

Cryptographic scheme requiring multiple parties to approve a decision.

P

Pipeline

Sequential processing flow through SCBE’s 14 layers.

Poincare Ball

Mathematical model for hyperbolic space, used in trust calculations.

Policy

Rules defining acceptable behaviors, thresholds, and decision criteria.

Post-Quantum Cryptography (PQC)

Cryptographic algorithms resistant to attacks by quantum computers.

Q

QUARANTINE

Decision outcome holding an AI agent request for human review. Issued when trust score is between 0.30 and 0.70.

Quantum-Resistant

Property of being secure against both classical and quantum computer attacks.

R

Risk Level

Categorization of request danger: LOW (0.70-1.00), MEDIUM (0.30-0.70), HIGH (0.00-0.30).

RWP (Read-Write-Permit)

Policy framework defining what AI agents can read, write, and execute.

S

SCBE

Secure Cryptographic Behavioral Envelope - the core governance framework.

Sentinel

Monitoring component that watches for anomalies and triggers alerts.

Steward

Administrative component managing policy updates and system configuration.

T

Trust Score

Numerical value (0.0 to 1.0) quantifying confidence in an AI agent request.

Trust Engine

Component computing trust scores using hyperbolic geometry and behavioral analysis.

V

Validator

Node participating in consensus decisions. Minimum 3 validators required (2f+1 where f=1).

Z

Zero Trust

Security model assuming no implicit trust; every request must be verified.

ZBase32

Encoding scheme used for human-readable identifiers in SCBE.

Mathematical Notation

Symbol	Meaning
d(u,v)	Hyperbolic distance between points u and v
		x	Euclidean norm of vector x
f	Maximum number of faulty validators
2f+1	Minimum validators for BFT consensus
arcosh	Inverse hyperbolic cosine