ThreatBook Acquires Open-Source AI Penetration Testing Platform CyberStrikeAI

Singapore and Hong Kong, 30 September 2026 — ThreatBook, the agentic security company, has acquired CyberStrikeAI, the open-source AI penetration testing platform. Written in Go for high performance, CyberStrikeAI lets security teams run continuous, automated assessments inside a controlled environment — surfacing the vulnerabilities that conventional testing misses so organizations can make faster, better-informed decisions about risk.

Since its launch in late 2025, CyberStrikeAI has attracted more than 7,000 stars on GitHub, while public internet-scanning data points to more than 3,000 instances in use worldwide — making it one of the most widely used AI penetration testing tools globally.

ThreatBook will keep the open-source edition of CyberStrikeAI available and will launch an Enterprise Edition alongside it. The Enterprise Edition can be deployed entirely on-premises, keeps all data local, and captures every agent action in a full audit trail with permission controls. ThreatBook will also add further safeguards to the open-source edition to prevent the technology from being misused.

Leveling the playing field for defenders

Feng XUE, Founder and CEO of ThreatBook, said: “AI attacks are now evolving far faster than anyone could have imagined. The idea that attackers need time to plan and stage an attack is outdated — and simply untrue. Automated reconnaissance, vulnerability discovery and exploitation have already eaten into the response window that defenders once relied on.”

“These capabilities should not sit only in the hands of cybercriminals, APT groups and saboteurs. They belong with the defenders who are committed to building a safer world — leveling the playing field and letting defenders think and act the way attackers do. CyberStrikeAI is mission-critical technology that helps security teams get there.”

Key CyberStrikeAI capabilities include: one-line deployment, for instant, hassle-free setup; native Model Context Protocol (MCP) orchestration, using standardized tool invocation that can be extended freely through YAML files; full attack-chain coverage, with more than 100 built-in tool templates spanning reconnaissance through exploitation; visual attack-chain tracking, so that every step the agent takes is observable; a multi-model decision engine that lets users plug in their preferred large language model (LLM) as the reasoning core; and natural language in, report out — describe a target, and receive a structured security assessment report.

ThreatBook will continue to invest in developing both editions of CyberStrikeAI, making the platform’s autonomous penetration testing capabilities more reliable, more controllable and easier to use. A trial version of the Enterprise Edition is available in Mainland China effective immediately, and is expected to reach the global market in the fourth quarter of 2026.