Developing Story
Autonomous AI Agents for Offensive Security Testing
CodeWall Ltd., a London startup, launched with $4 million in funding to offer autonomous AI agents that simulate real-world hacker attacks on enterprises like banks, part of a broader emerging category of AI-driven offensive security testing.
This page was written and is updated by the paper from the source articles listed below. Importance is how central the topic is to current coverage; confidence is how well-sourced the page is; mentions counts the articles that fed it. How the wiki works →
A new wave of startups is commercializing autonomous AI agents designed to attack enterprises the way outside hackers would, as a form of continuous offensive security testing. London-based CodeWall Ltd. launched with $4 million in funding to offer enterprises, such as banks, AI agents that begin engagements with no credentials or inside knowledge of the target, deploying several specialized agents for different attack phases (SiliconANGLE, October 14).
### Why it matters
This represents an emerging commercial category — autonomous red-teaming-as-a-service — with implications for enterprise cybersecurity budgets, regulatory compliance (e.g., financial sector security mandates), and liability questions around AI systems that actively probe production environments. Attorneys advising security vendors or enterprise clients should track how liability, authorization, and data-handling frameworks develop for this category as more entrants emerge.
### Key facts
- CodeWall Ltd. is based in London and launched October 14, 2026 (SiliconANGLE, October 14).
- Raised $4 million in new funding (SiliconANGLE, October 14).
- Targets enterprises such as banks with autonomous AI agents simulating external attackers (SiliconANGLE, October 14).
- Engagements reportedly begin with zero credentials or inside knowledge of the target (SiliconANGLE, October 14).
- Uses multiple specialized agents, each built for a different attack phase (SiliconANGLE, October 14).