FireCompass Secures USPTO Patent for Continuous Automated Red Teaming (CART)

FireCompass has recently secured a patent from the United States Patent and Trademark Office (USPTO) for their innovative approach to Continuous Automated Red Teaming (CART) in Organizational Networks. This milestone is set to reshape the landscape of cybersecurity, particularly in the realms of automated penetration testing and red teaming.

The minds behind this revolutionary technology are Bikash Barai (CEO & Co-Founder, FireCompass), Arnab Chattopadhayay (Co-Founder & VP Of Emerging Research, FireCompass) and Jitendra Chauhan (Head of R&D, FireCompass). They have successfully combined Reinforcement Learning and AI/ML algorithms to transform External Attack Surface Management (EASM) and automated pen testing.

Why CART needs a new foundation

Attack surfaces are expanding faster than manual testers and point-in-time exercises can keep up, while attackers iterate continuously and string seemingly minor exposures into working breach paths. Traditional quarterly pen tests and manual red team operations leave gaps between windows of testing, and playbook-based automation struggles when assets change daily or hourly.​

To materially reduce breach likelihood, programs need a system that can continuously discover external exposures, prioritize likely chains, safely validate them, and learn from outcomes—without depending on a single central planner or fragile playbooks.​

What the FireCompass patent covers

The USPTO-granted invention describes “a system and method to perform automated red teaming in an organizational network,” built from coordinated subsystems that discover attack surfaces, identify attack frontiers, prioritize and emulate multi-stage attacks, determine attack paths, and learn continuously to generate an AI-based security model.​
Crucially, it discards centralized orchestration: each attack subsystem self-launches based on predefined conditions and posts outcomes to a shared layer, enabling “automated choreography” that scales and adapts when the environment changes.
What’s novel vs. prior

• No attack-graph compute bottleneck: The system avoids full attack-graph generation for every state, instead using a frontier-driven, state-aware process that scales better as surfaces change.​

• No central controller dependency: Subsystems self-launch on conditions and publish outcomes, mitigating latency and control-plane fragility.​

• Continuous learning loop: The model updates with every emulation result, improving frontier selection, pathing, and risk ranking over time.​

>>Outpace Attackers With AI-Based Automated Penetration Testing

FireCompass CART Engine ( USPTO Patent)

• Secure data gathering → Collects minimal input (e.g., domain/IP/email) plus an exit criterion, then pulls data from multiple sources into a structured graph.​ Z‑Graph indexer and message-board driven updates that build the working state/graph from minimal inputs plus exit criteria.​

• Attack surface determination → Automatically identifying least-secure points that form the exploitable external surface.​

• Attack frontier identification → Enumerating candidate attacks “given the current state of Z‑Graph” for each surface.​

• Prioritization → Strategies including anomaly and state‑change detection to decide which frontiers to try next.​

• Emulation and simulation → Safely executing multi‑stage attacks on prioritized frontiers to generate empirical results.​

• Attack path determination → Converting successful steps into end‑to‑end chains reflecting real lateral movement and privilege escalation.​

• Learning subsystem and AI‑based model → Ingesting outcomes and using multiple learning techniques (including reinforcement‑style trials) to correlate patterns and assign per‑path risk values; this is echoed in FireCompass’ CART materials emphasizing validated exploits and attack‑path visualization.

Exploit-verified risk: attack paths, not alerts

Instead of listing potential issues, the system validates end-to-end attack chains and assigns per-path risk, letting teams fix what’s actually exploitable and de‑prioritize noise.​
Active, safe execution and attack-tree chaining are positioned to minimize false positives compared to discovery-only scanners, with FireCompass materials highlighting live exploit proofs and path visualization.​

Zero guesswork: near‑zero false positives through active validation

CART integrates reconnaissance with safe exploitation to confirm whether a weakness is real in the current environment, reducing alert fatigue and wasted cycles on non‑issues.​
FireCompass guidance contrasts exploit‑validated findings with scanner noise and emphasizes repeated, event-triggered replays to keep validation current as assets change.​

How does this improve CART and automated pen testing for CISOs?

• Path-first risk reduction: Instead of tallying isolated findings, the system ranks breach likelihood by end-to-end attack paths and their learned patterns.​

• Faster validation: Safe, autonomous execution demonstrates exploitability across chains, compressing the time from discovery to proof and remediation.​ Non-viable attack paths have reduced risks associated and only exploitable vulnerabilities are surfaced.

• Continuous coverage: As new assets and behaviors appear, the frontier and priorities update automatically, sustaining pressure on changing attack surfaces.

CISO outcomes: Provable reduction in exploitable risk

Continuous Automated Red Teaming only moves the needle when it validates real chains and keeps pace with change; the patented approach achieves this by replacing fragile orchestration with scalable choreography and a learning loop that prioritizes what to validate next.​ For CISOs, the net is exploit‑verified, path‑level risk you can measure & act upon—zero false positives, faster proofs, and clearer remediation focus tied to actual breach paths.

• Reduction in viable attack paths: Measure the number and criticality of validated chains eliminated over time to quantify the reduction in breach likelihood.​

• Discovery-to-validation time: Track the interval from new exposure detection to safe proof to prove operational velocity gains.​

• Coverage and cadence: Monitor the percentage of external assets under continuous test and the frequency per asset class to ensure incrementally better security posture.

US Patent Announcement @DSCI, AISS 2023

-> (Request Demo) 10 X Pen Testing Frequency & 100% Asset Coverage

Outpace Attackers With AI-Based Automate Penetration Testing With FireCompass:

FireCompass is a single platform for AI-Powered Continuous Automated Red Teaming (CART), Pen Testing & NextGen Attack Surface Management

>>FireCompass Free Trial