CodeWall Raises $4M for Autonomous Attack Agents

CodeWall has emerged from stealth with $4 million to build autonomous offensive-security agents that search for exploitable weaknesses in customer-authorised systems.

What happened

The London-based startup raised pre-seed funding led by Osney Capital, with participation from InReach Ventures, Antler and angel investors. CodeWall's agents attempt to penetrate systems without being given credentials or detailed advance knowledge, exploring different attack paths and sharing discoveries through a common investigation graph. The product is designed to provide evidence showing how a compromise could occur rather than simply list theoretical vulnerabilities. Testing is restricted to environments where customers have explicitly authorised the activity.

Why it matters

Traditional penetration testing is often periodic and labour-intensive. AI can make adversarial testing more continuous, helping security teams identify combinations of weaknesses before real attackers exploit them. The value lies in showing which exposures are actually exploitable rather than adding more alerts.

The bigger picture

Autonomous offensive security creates its own safety requirements. Strong controls, audit logs and strict scoping are essential because the same capabilities used for authorised testing could be dangerous outside controlled environments. CodeWall's credibility will depend on both technical effectiveness and responsible operating boundaries.