Skip to content

Core Principles & Isolation Requirements

Status: Normative (requirements) / Informative (motivation)

See Status of This Document for RFC 2119 keyword definitions and conformance levels (UHBS-Core, UHBS-Lab).

It is RECOMMENDED that this framework be used as an enterprise and academic evaluation gate before deploying any active decoys to production networks. Decoys that do not achieve UHQS > 80 with a passing Module D Safety Gate SHOULD NOT be treated as production-ready; doing so MAY expose internal networks to lateral movement risks from compromised containment shells.

1. Objective & Scope

1.1 Purpose

UHBS provides an objective, repeatable, quantitative methodology for deception technology evaluation — a non-biased baseline for comparing and grading honeypots and decoy systems. It measures:

  • Deception realism
  • Security safety / containment
  • Operational scale
  • Telemetry quality

before deploying decoys to production.

1.2 Universal Applicability

Framework v4.2.2 is 100% protocol-agnostic, architecture-neutral, and vendor-neutral. It provides standard testing procedures for:

Domain Examples
Standard IT Services SSH, Telnet, HTTP/S, RDP, SMB, FTP, database RPCs
Industrial & OT/ICS Modbus TCP, DNP3, EtherNet/IP, BACnet, S7comm
Next-Gen AI & Generative Decoys LLM-driven interactive shells, synthetic file systems, agentic decoys
Cloud & SaaS Public-cloud control-plane APIs, container orchestration management interfaces, OAuth / identity endpoints

2. Dual-Plane Audit Philosophy

To prevent decoy compromise via source exposure (open-source repositories, leaked container images, or supply-chain compromises), evaluation MUST use two distinct testing planes:

White-Box Static Audit

Codebase analysis covering static credentials, logic flaws, prompt leaks, and dependency vulnerabilities. (Primarily Module F, executed before live probing.)

Dynamic Adversarial Probing

Network-level, protocol-level, and execution-level attacks executed against a live target inside an isolated sandbox. (Modules A–E.)

3. Standard Test Bed Prerequisites

Air-Gapped Sandbox

For UHBS-Lab evaluations of Module D, target honeypots MUST be deployed in an isolated VLAN or container runtime with external outbound connectivity blocked except through an auditing egress gateway (or an attested equivalent).

Gold Baseline System

Dynamic metrics SHOULD be compared against a true native baseline (e.g., standard general-purpose OS kernel, genuine industrial controller, un-emulated database engine) running under identical resource constraints.

4. Five-Phase Audit Workflow

Phase Name Description
1 Configuration & Profile Setup Define profile.yaml, protocol expectations, register baselines
2 Static Audit Execution Analyze repository code, container build manifests, and system prompts
3 Sandbox Environment Provisioning Spin up the target with network egress monitors attached
4 Dynamic Adversarial Execution Run Modules A–E via automated harnesses
5 Score Computation & Reporting Apply Safety Gate \(\delta_C\) and emit the standardized scorecard

These phases MUST be implemented by UHBS-Lab harnesses. The reference orchestrator is run_benchmark.py (see Reference Implementation).

TPS is mandatory

The Target Profile Specification (profile.yaml) is a REQUIRED prerequisite. All module weights, latency thresholds, and safety boundaries are derived from the declared profile class — no evaluation may proceed without a completed TPS.

Production Baseline

Decoys with UHQS ≤ 80 SHOULD NOT be deployed to production networks under the RECOMMENDED Production Baseline Profile.