Shackel believes that a market in organs would be a good thing, even for the poor people who are most likely to be the ones selling organs, because it allows the buyer to gain a kidney, greatly extending their life, and the seller to make a lot of financial headway. He describes this as a win-win situation. But believing that this is a fair trade is exactly which leads to there being obligations to trade on the market, as Lewis's framework shows.
A quick, simplified gloss on Lewisian conventions: there is a convention to do C (and we should do C rather than not) when almost everybody expects people in the relevant circumstances to do C (and expects people to expect doing C), prefer doing C given this expectation, and would prefer some other action C* if almost everybody expected to do C* instead. This gives rise to straight-forward obligations because when I act against these expectations, I frustrate everybody's ability to co-ordinate towards mutually preferred ends. This is all pure game theory to this point, so I don't expect you have much to object to here. So, for the organ market case, it would go: for any legitimate market, I expect that trading occurs on that market and I prefer there being trade on a market over no trade (otherwise there wouldn't really be a market), which covers the 1st and 2nd conditions. If there were no market, or if it were banned, I prefer no trade on it, like I do for the market for hard drugs or illicit weapons (the 3rd condition). Because the three conditions are met, there is a convention to trade on the organ market. Thus, if I am in the relevant circumstances -- needing money and having something to trade on the market -- there is a convention to make use of the market, and an obligation to do so, otherwise I am arbitrarily frustrating people's attempts to reach the type of scenario Shackel describes as win-win. But this means that a market in organs leads to at least some obligations.
Shackel is committed to this result exactly because he describes someone improving their economic condition through organ trade as a win-win situation. To deny that somebody would prefer trade to non-trade (denying the 2nd condition I outlined above, and the only plausible response) would mean denying that such an exchange, even at full market value, would be fair, and that would be to deny the good of the organ market. Shackel makes the claim that we have second-order norms on the fairness of expectations which would stop exploitation, and such norms are certainly present and in effect. But for there to be such norms describing even trade at full market value as an unfair expectation would mean that there are constraints on the fairness of trade which goes beyond what determines fair market value. Those of us who are more skeptical of the goodness of such a market have that option available to us, but Shackel does not.