Governing what an autonomous agent is allowed to spend

How do you control what an autonomous AI agent is allowed to pay, trade, or deploy?

Once an agent holds a key it holds spending power, and the question stops being what it should do and becomes what it is permitted to do. The answer is a policy layer in front of the key: every payment, swap, and deploy checked against operator rules before it is signed, denied on the spot when it breaks one, and the decision anchored where nobody can revise it afterwards. AgentOps is my build of that on Casper, one organization budget governing x402 payments, CSPR.trade swaps, and direct deploys through an on-chain GuardRegistry, exposed to agents over MCP.