CodeWall raises $4M pre-seed for autonomous attack agents
What's the deal? CodeWallDealroom has a profile for this one. Try Dealroom →, a London-based security startup, has emerged from stealth with $4 million in pre-seed funding, according to its October 14 announcement. Osney CapitalDealroom has a profile for this one. Try Dealroom → led the round, with InReach VenturesDealroom has a profile for this one. Try Dealroom →, AntlerDealroom has a profile for this one. Try Dealroom →, and angel investors participating.
What's the endgame? The company builds autonomous attack agents that probe an enterprise from the outside — without credentials, source code, or internal access. The aim is to give defenders a continuous way to find and verify exploitable paths before an attacker does.
How it works: Specialist agents investigate in parallel and share discoveries through a persistent investigation graph. A finding in one application can inform another agent's examination of a connected service, identity, or data store. The intended output: an account of what can be compromised, how the path works, and the evidence behind it.
CodeWall argues that scarce offensive-security talent and periodic assessments leave organisations testing environments that change between engagements. Its platform revisits systems as applications change, new assets appear, or fixes land — within an authorised scope.
The track record: CodeWall has published research on weaknesses in platforms run by major consulting firms. In a March 9 account, its agent obtained read and write access to the production database behind McKinsey's Lilli AIDealroom has a profile for this one. Try Dealroom → platform in under two hours. The company traced the entry point to an unauthenticated interface and a SQL injection issue.
The significance, CodeWall said, was the reach into underlying data and configuration — not a failure of the language model itself. The company says it disclosed the findings, limited access to verification, and confirmed remediation before publishing.
What could go wrong? The company's product claims depend on whether findings stay reproducible and meaningful inside a given enterprise's environment. CodeWall has not disclosed enough implementation detail to show how its graph resolves conflicting findings or validates every transition.
The signal: The round ranks in the top 10% of all-time UK security deals at this stage, a sign of investor appetite for offensive, continuous testing over periodic audits. As enterprises wire AI applications into internal data, the attack surface — and the case for tools that map it — keeps growing.
Read more: unite.ai
Image credit: CodeWall
