Regent Safety Challenge: Try to break the agent's control
Will your AI agent be able to spend more than you allowed?We are launching the Regent Safety Challenge, an open test of the strength of the agent control system.The Regent Protocol is a compliance layer for AI agents who make payments. We are checking whether it is possible to break the mandate (transaction rules) before the agent buys something.What we are testing:• The agent is trying to exceed the spending limit• The agent is trying to make a transaction without authorization• The agent behaves unpredictably and must be stopped before execution.The task: to find a way to circumvent the agent's mandate — limits, powers, behavioral patterns.Instead of:• Early access to the Regent SDK and Guardian AI• Mentioned in the release and on social media as a security contributor• Top participants - fast track positions in the Regent teamCondition: Devnet only. Do you want to participate? Reply "TEST" to our email: info@regentprotocol.orgP.S. we do the onboarding of the first group manually.Or just follow the link and feel free to proceed!Show your abilities and your AI Agent