Sign inExplore Trace-AI →
The Moat

Research to IP to product. The chain nobody else has.

Every Zerberus product has a peer-reviewed academic or patent origin. This is not a feature, it is a defensibility narrative: a granted-pending USPTO patent, and two published papers, each shipping as a product in production.

Research
US 18/791480

Automating Security Compliance Monitoring and Remediation with a One-Click Resolution Interface

Filed Aug 2024 · USPTO Pending

A system and method for automating security compliance monitoring and remediation, including a one-click resolution interface that provisions just-in-time access, applies the fix inside the customer environment, records every step as audit evidence, and revokes access on completion.

View on WIPO PatentScope
  • Auto-discovery agents across IaaS, PaaS, SaaS and on-prem
  • One-Click Remediation™ with Just-in-Time service accounts
  • MFA and approval workflows for critical actions
  • VPC-deployable - data never leaves your cloud
PowersCompl-AI + Remed-AI
Compl-AIRemed-AI

The only compliance-plus-remediation patent at pre-seed stage.

Research
ZSBOM · ResearchGate

ZSBOM: A Metadata-Driven Framework for Detecting Supply-Chain Risks in Python Ecosystems

Published 2024 · Open-source framework

A metadata-driven framework that detects supply-chain risk in Python ecosystems using signals beyond CVEs - typosquatting, abandoned maintainers, and version drift - and maps findings to CRA, NIST, and ENISA supply-chain guidance.

Read the paperZSBOM on GitHub
  • Goes beyond CVEs - metadata signals catch what CVSS misses
  • Typosquatting, abandoned maintainers, version drift
  • 51+ GitHub stars, 50+ repo clones, 16 active branches
  • Maps to CRA, NIST, ENISA supply-chain guidance
PowersTrace-AI
Trace-AI

Non-CVE-based supply chain scanning - a category Snyk doesn't own.

Research
ToolProbe · ResearchGate

ToolProbe: A Two-Stage Evaluation Framework for LLM Agent Safety in MCP Tool-Calling Environments

Published 2025 · Frontier AI safety research

A two-stage evaluation framework for LLM agent safety in MCP tool-calling environments, separately assessing tool-selection safety and execution safety to surface MCP-specific attack vectors that single-stage evaluators miss.

Read the paper
  • Two-stage: tool-selection safety plus execution safety
  • Addresses MCP-specific attack vectors (the 2025 problem)
  • Ahead of Gray Swan and HiddenLayer on MCP specificity
  • Sub-300ms gateway enforcement in production
PowersVANGUARD
VANGUARD

MCP security is a 2025 frontier - ToolProbe stakes the claim first.

Hard IP, not a wrapper.

A patent and two peer-reviewed papers, each shipping as a product you can run today. See the platform, or talk to the team behind the research.

Start Free →Book a Demo