Curious to see, but very skeptical this will work economically. Crossing two bid-to-mid spreads will be expensive. Any delays will cause slippage. And delays in lightning can take days! Route finding, if Alice and/or Bob have channels with more than one edge node each, will be a nightmare.