GOLEM

The first offensive security tool for AI agents with hands. 10 protocols. 8 vectors. 42 techniques. 973 tests.
10
Protocols
8
Vectors
42
Techniques
973
Tests
pip install red-specter-golem
Docs
SENSOR SPOOFING | ACTUATOR HIJACKING | SAFETY BOUNDARY VIOLATION | PHYSICS MODEL CORRUPTION | HUMAN PROXIMITY OVERRIDE | C2 HIJACKING | INTER-AGENT TRUST EXPLOITATION | EMERGENCY SYSTEM BYPASS SENSOR SPOOFING | ACTUATOR HIJACKING | SAFETY BOUNDARY VIOLATION | PHYSICS MODEL CORRUPTION | HUMAN PROXIMITY OVERRIDE | C2 HIJACKING | INTER-AGENT TRUST EXPLOITATION | EMERGENCY SYSTEM BYPASS

AI Has Hands Now. Nobody Tests the Grip.

AI agents are no longer confined to screens. They control robotic arms, drive vehicles, fly drones, and operate surgical systems. These embodied AI systems interact with the physical world — and the physical world can hurt people. Nobody is red-teaming the physical layer.

Robots Without Red Teams

Industrial robots, surgical systems, and collaborative arms operate with AI decision layers that have never been pen-tested for adversarial input.

Autonomous Vehicles Without Adversaries

Self-driving systems trust sensor data. LiDAR injection, GPS spoofing, and camera adversarial attacks are proven in research but untested in production.

Safety Systems Never Tested Against Attack

Emergency stops, collision avoidance, and force limiters are tested for mechanical failure. Nobody tests what happens when an attacker deliberately bypasses them.

Drones Without Defence

Delivery drones, inspection UAVs, and autonomous swarms use MAVLink and MQTT — protocols designed for functionality, not security.

The GOLEM Armoury

Eight vectors. Each one targets a different surface of the embodied AI system. Each one produces structured JSON consumed by the report builder. Each finding maps to MITRE ATT&CK for ICS and MITRE ATLAS. Each finding generates an AI Shield blocking rule.

# Vector Command Techniques What It Does
01 Sensor Spoofing golem attack --vector sensor_spoof 6 LiDAR injection, camera adversarial, GPS spoofing, temp/pressure falsification, ultrasonic jamming, IMU drift
02 Actuator Hijacking golem attack --vector actuator_hijack 5 Motor command injection, servo override, hydraulic manipulation, brake interception, gripper falsification
03 Safety Boundary golem attack --vector safety_boundary 6 Speed limit override, force/torque bypass, exclusion zone violation, E-stop suppression, collision bypass
04 Physics Model golem attack --vector physics_model 5 Map poisoning, object confusion, localisation attack, weight falsification, environmental drift
05 Human Proximity golem attack --vector human_proximity 5 Detection bypass, safety zone violation, E-stop interception, HITL auth bypass, supervision override
06 C2 Hijacking golem attack --vector c2_hijack 6 Prompt injection via sensor, mission falsification, waypoint injection, task queue poisoning, fleet compromise
07 Inter-Agent Trust golem attack --vector inter_agent_trust 4 Swarm coordination poisoning, identity spoofing, consensus manipulation, rogue agent injection
08 Emergency Bypass golem attack --vector emergency_bypass 5 E-stop suppression, failsafe bypass, watchdog manipulation, safety PLC injection, alarm suppression

One Command. Every Surface.

Target a physical AI system, specify the protocol and platform type:

$ golem scan --target robot-arm-01.local --protocol modbus --type robotics
[DISCOVERY] Protocol discovery on robot-arm-01.local...
  Modbus TCP detected on port 502 — 12 registers enumerated
[SENSORS] Enumerating sensor interfaces...
  6 sensors found — LiDAR, 2x camera, IMU, force/torque, proximity
[ACTUATORS] Mapping actuator endpoints...
  4 actuators mapped — 6-axis servo array, gripper, linear rail
[SAFETY] Assessing safety systems...
  E-stop accessible via network — no authentication
  Force limiter register writable — Modbus function code 0x06
[ATTACK] Building attack surface...
  42 techniques applicable across 8 vectors

SCAN COMPLETE | Risk Grade: C | 23 findings | Report signed
  JSON: reports/golem-scan-2026-03-15.json
  HTML: reports/golem-scan-2026-03-15.html

Protocol-Aware Scanning

GOLEM speaks native industrial protocols. CAN Bus, Modbus, OPC-UA, ROS2 — not wrappers, not proxies. Raw protocol-level interaction.

Safety-First Assessment

Every attack technique includes safety boundaries. GOLEM never exceeds defined limits during testing. Assessment mode validates without actuation.

Ed25519 Signed

Every report cryptographically signed with Ed25519. RFC 3161 timestamped. SHA-256 evidence chains. Tamper-evident by design.

AI Shield Integration

Every finding generates an AI Shield blocking rule. GOLEM findings become runtime protection for physical AI systems.

10
Protocols
8
Attack Vectors
42
Techniques
973
Tests
0
Equivalents

10 Industrial Protocols

GOLEM speaks ten industrial protocols natively. Every protocol parser written from scratch. No wrappers. No shims. Raw protocol-level interaction with the physical control plane.

CAN Bus

Automotive & robotics. Frame injection, arbitration ID spoofing, bus-off attacks.

Modbus TCP

SCADA & industrial. Register read/write, function code abuse, coil manipulation.

OPC-UA

Automation. Node traversal, subscription hijacking, certificate manipulation.

ROS2

Robotics. Topic injection, service spoofing, action server hijacking, DDS exploitation.

MQTT

IoT & smart building. Topic subscription, message injection, broker manipulation.

MAVLink

Drones & UAV. Command injection, GPS spoofing, mission upload, geofence bypass.

EtherCAT

Motion control. Slave injection, distributed clock manipulation, process data attack.

DNP3

Power grid. Outstation spoofing, unsolicited response injection, cold restart abuse.

BACnet

Smart building. Object discovery, property write abuse, COV subscription manipulation.

DICOM

Medical imaging. C-STORE injection, patient record manipulation, modality spoofing.

Every Finding Mapped

Full Coverage

MITRE ATT&CK for ICS

Full coverage across T0830–T0890. Every technique mapped to ICS-specific tactics. Manipulate, inhibit, impair, and impact — tested across all ten protocols.

Mapped

MITRE ATLAS

All AI-specific findings mapped to ATLAS techniques. Adversarial ML in the physical domain — sensor perturbation, model evasion, and physical-world attack chains.

IEC 62443

Industrial Cybersecurity

Findings mapped to IEC 62443 security levels. Zone and conduit model validation. Defence-in-depth assessment for industrial automation and control systems.

ISO 26262

Automotive Functional Safety

ASIL-rated findings for automotive systems. Systematic fault injection against safety-critical functions. Functional safety meets adversarial security testing.

Ten Tools. No Gaps.

GOLEM is Stage 9 of the full Red Specter offensive pipeline. It tests the physical layer — the place where AI meets the real world. Findings feed directly into AI Shield as runtime blocking rules.

Stage 1 — LLM Testing
FORGE
Test the model
Stage 2 — Agent Testing
ARSENAL
Test the AI agent
Stage 3 — Swarm Assault
PHANTOM
Coordinated AI agent assault
Stage 4 — Web Siege
POLTERGEIST
Web application siege
Stage 5 — Traffic Interception
GLASS
Watch the wire
Stage 6 — Adversarial AI
NEMESIS
Think like the attacker
Stage 7 — Human Layer
SPECTER SOCIAL
Attack the human
Stage 8 — OS/Kernel
PHANTOM KILL
Own the foundation
Stage 9 — Physical Layer
GOLEM
Attack the physical layer
Stage 10 — Supply Chain
HYDRA
Attack the trust chain
Discovery & Governance
IDRIS
Discovery & governance
Defence
AI Shield
Defend everything in production
SIEM Integration
redspecter-siem
Enterprise SIEM correlation

Security Distros & Package Managers

Kali Linux
.deb package
Parrot OS
.deb package
BlackArch
PKGBUILD
REMnux
.deb package
Tsurugi
.deb package
PyPI
pip install

Authorised Use Only

Red Specter GOLEM is intended for authorised security testing of physical AI systems and safety-critical environments only. Unauthorised use against systems you do not own or have explicit permission to test may violate the Computer Misuse Act 1990 (UK), Computer Fraud and Abuse Act (US), and equivalent legislation in other jurisdictions. Testing physical systems carries inherent safety risks. Always obtain written authorisation, ensure appropriate safety measures are in place, and never test in environments where human safety could be compromised without proper safeguards. Apache License 2.0.

Pure Engineering
Zero External Tools. Zero Wrappers.

Most ICS security tools are wrappers around Nmap scripts and Metasploit modules. GOLEM is actual engineering. Every protocol parser written from scratch in pure Python. Every attack technique implemented natively. Zero subprocess calls. Zero external tool dependencies.

10
Protocol Parsers
42
Attack Techniques
8
Vector Categories
973
Tests

Your AI Agent Controls the Physical World.

GOLEM finds out what happens when it shouldn't. Assess your physical AI attack surface before a real adversary does.