ChainReactors' Open-Source Engine Is Rewriting the AI Attack Playbook

In the high-stakes world of cybersecurity, the most dangerous vulnerabilities are often the ones you don't see coming. They are not single flaws, but chains of them, strung together by a human attacker's creativity to escalate a minor foothold into a catastrophic breach. A startup in Lisbon is betting that an AI, not a human, should be the one searching for these hidden paths.

ChainReactors, founded in 2021, is building what it calls an AI-native offensive security platform. Its core proposition is a shift from scripted vulnerability scanners to autonomous agents that can reason about a network, discover attack sequences, and adapt their approach in real time. The company's initial wedge is an open-source red-team scanning engine named gogo, which has drawn over 2,000 stars on GitHub, signaling early adoption from the security community [Dealroom News, Unknown]. This open-source traction is the foundation for a commercial play that aims to redefine how enterprises test their own defenses.

The bet on autonomous attack planning

The company's ambition moves beyond automating known exploits. Its research, including a project called ChainReactor developed in an academic context, models privilege escalation as an AI planning problem [GitHub - ucsb-seclab/chainreactor, Unknown]. The system treats a network as a shared state space, where the AI agent is not given a preset workflow but must independently search for viable sequences of actions to reach a goal, like gaining administrative access. This approach, the company claims, can rediscover known exploit chains and potentially identify novel ones that human testers might miss [GitHub - ucsb-seclab/chainreactor, Unknown].

ChainReactors is productizing this research. Its commercial platform, Cairn, alongside tools like an AI-driven agent named aiscan and a command-and-control framework called IoM, is designed to bring this autonomous planning capability into the hands of penetration testers and red teams [ChainReactors, retrieved 2024]. The company reports its tools have already been used to find over 180 severe and 740 high-risk vulnerabilities in real network environments [ChainReactors, retrieved 2024].

An open-source first go-to-market

ChainReactors is pursuing a classic product-led growth strategy in a field dominated by large, closed-source vendors. By open-sourcing the core kernels of its technology, like the gogo scanner, it builds credibility and a user base within the security practitioner community. The commercial layer then adds enterprise-grade features, management, and support. This model is reflected in its funding: the startup has raised a total of $449,000 in seed capital from investors including DigiBuild and Morgan Creek Capital Management [Dealroom News, Unknown].

The founding team brings a blend of deep technical and unconventional backgrounds. CEO Nate Petre holds a PhD and has worked on projects ranging from reef-building to NASA-sponsored research on seaweed surfboards [Conversation with Nate Petre of Chainreactor - FoodTech Weekly, Unknown]. Co-founder Ke, born in 1998, represents a younger generation of security talent [Dealroom News, Unknown]. The team's public record shows experience in biotechnology and additive manufacturing, suggesting a cross-disciplinary approach to problem-solving [Chainreactor.bio, retrieved 2026].

The competitive and technical hurdles

For all its promise, the path for an AI-driven offensive security platform is fraught with technical and market challenges. The field of automated penetration testing is crowded, and convincing security chiefs to trust an autonomous AI with offensive operations is a significant leap. The technology must prove it is not just a clever scanner but a reliable partner that can operate within strict ethical and operational boundaries.

The company's current traction is promising but early. While a GitHub star count shows developer interest, it does not directly translate into enterprise revenue. Furthermore, the leap from a research project demonstrating capability on synthetic virtual machines to a robust, scalable product for complex enterprise networks is substantial [GitHub - ucsb-seclab/chainreactor, Unknown].

Key questions ChainReactors must answer in the next 12-18 months include:

• Commercial validation. Converting open-source users into paying enterprise customers, particularly for its flagship Cairn platform.
• Technical robustness. Proving the AI agent's effectiveness and safety in diverse, production-grade network environments beyond controlled tests.
• Regulatory posture. Navigating the complex ethical and legal landscape surrounding autonomous offensive security tools, even when used defensively.

What penetration testing looks like today

The standard of care in offensive security today is a hybrid, human-intensive process. It typically involves teams of highly skilled ethical hackers using a combination of manual techniques and automated scanning tools from vendors like Tenable, Rapid7, or open-source frameworks. These tools excel at finding known vulnerabilities (CVEs) but often struggle with chaining them together or reasoning about unique network architectures. The result is a process that is both expensive, due to the scarcity of expert talent, and potentially incomplete, as it relies on human intuition to connect disparate dots. ChainReactors is aiming at this gap, proposing a future where AI assistants can augment human red teams, continuously probing for attack paths and ensuring defenses are tested against increasingly sophisticated, automated threats.

The company's next milestones will likely involve scaling its commercial offerings and securing a larger round of funding to accelerate enterprise sales. With its open-source engine already in the wild, ChainReactors has planted its flag. The harder task, as always in clinical AI and security, is moving from a compelling demo to a trusted, indispensable component of the security stack.

Sources

1. [Dealroom News, Unknown] Chainreaction raises $449K seed to rewrite how security is priced | https://dealroom.co/news/165434-chainreaction-raises-449k-seed-to-rewrite-how-security-is-priced/
2. [ChainReactors, retrieved 2024] ChainReactors, AI + 进攻性安全 | https://www.chainreactors.ai/
3. [GitHub - ucsb-seclab/chainreactor, Unknown] ChainReactor is a research project that leverages AI planning to discover exploitation chains for privilege escalation on Unix systems. | https://github.com/ucsb-seclab/chainreactor
4. [Conversation with Nate Petre of Chainreactor - FoodTech Weekly, Unknown] Conversation with Nate Petre of Chainreactor | https://foodtechweekly.com/substack-posts/conversation-with-nate-petre-of-chainreactor
5. [Chainreactor.bio, retrieved 2026] Chainreactor, The infrastructure for industrial biology | https://www.chainreactor.bio/
6. [Open-Source Security Startup "Chain Reaction" Secures Multi-Million RMB Seed Round, Rewrites Industry Pricing Power With Token-Based Billing, Unknown] Open-Source Security Startup "Chain Reaction" Secures Multi-Million RMB Seed Round, Rewrites Industry Pricing Power With Token-Based Billing | https://eu.36kr.com/en/p/4037822855058307