Deception technology is pretty interesting, but it seems like credential guard (and other ASR features) make it really hard to put legitimate-looking credentials in memory. Aside from canary tokens on disk, any interesting research or tricks here?