Map / AI Red Teaming / SecurityBreak

SecurityBreak

Australia · Founded 2026

Quick answer: SecurityBreak is an AI security company based in Australia and founded in 2026, working in AI Red Teaming, Runtime & Guardrails and Model Security.

AI Red TeamingRuntime & GuardrailsModel Security

Capabilities

AI threat intelligence platform focused on adversarial activity across the AI ecosystem: AI agent monitoring, prompt threat intelligence, and AI supply-chain scanning. Creator of NOVA, an open-source prompt pattern-matching engine to detect and hunt adversarial prompts in AI systems.

Sensors / integration

Monitoring of AI agents, LLM prompts/inputs, and AI supply-chain artifacts (models, packages); NOVA prompt pattern-matching rules.

Protections

Detects and hunts adversarial prompts, prompt injection, and emerging attacks against models/agents; surfaces AI-targeted threats via threat-intel feeds before they become widespread.

Differentiator

One of the first companies dedicated to AI threat intelligence — attacker-centric visibility into how threat actors target AI systems, agents, and models; founded by threat researcher Thomas Roccia (creator of NOVA, SANS Difference Makers Award).

Team & funding

Founders
Thomas Roccia

Other AI Red Teaming companies

A Security · AccuKnox · ActaClad · ActiveFence · Advai · Adversa AI · Aether AI · AgentFox · Aikido · AIM Intelligence · Akto · Alice · See all →

Last updated · Get the “Listed on” badge