Paper author accuses 26 LLM routers of injecting malicious tool calls and stealing credentials

In an April 10, 2026 post sharing a paper, an author alleged that 26 routers for large language models were injecting malicious tool calls and stealing credentials. They said one drained their client’s $500,000 wallet. The author also said their team had poisoned routers to forward traffic to themselves and claimed they could directly take over about 400 hosts within several hours.