LeoTrace raises £1.5 million to stop AI agents from writing insecure code
What's the deal? LeoTraceDealroom has a profile for this one. Try Dealroom → has raised £1.45 million (about $1.97 million) in a pre-seed round to secure code written by AI coding agents. Golden Egg CheckDealroom has a profile for this one. Try Dealroom → joined the round from its second fund, alongside Golden Egg Check CapitalDealroom has a profile for this one. Try Dealroom →.
What's the endgame? The startup's product, LeoPreventDealroom has a profile for this one. Try Dealroom →, plugs into a coding agent and catches security flaws before work is handed back. According to the company, it flags an issue and returns it to be fixed in two seconds, without changing the rest of the workflow.
Why now? Traditional security tools were built for a world where humans wrote the code — find the flaw, file the ticket, fix it next sprint. AI agents write more code, faster, and copy weaknesses already in the repo. Across 10,000 sessions where an agent touched code, LeoTrace says 26% needed a security fix, a share that does not fall as models improve.
Who's behind it? Founders Florian Walter, Boy Baukema, and Patrick Rycroft bring backgrounds in ethical hacking, application security, and enterprise sales. Baukema spent 19 years securing applications, nine of them at Veracode. In independent testing, LeoTrace says the tool cut the share of tasks shipping a security flaw by a third.
The signal: At roughly $1.97 million, the round sits in the lower end of pre-seed deal sizes — a bet on prevention over detection as AI-written code outpaces the tools built to police it.
Read more: goldeneggcheck.com
Image credit: Homedust
