To start, much of the proposed above is based on trust rather than rules defined in an SSC. There are no measures preventing the consumer-to-consumer platform from just taking the money and leaving. I would recommend building a 3-Party SSC from some of the 2-Party SSC's available. In these examples, there are fail-safes in place that disincentive foul play.
Ultimately, this proposed scenario seems to be a derivation of the 2-Party Multisignature Escrow Account with Time Lock and Recovery presented in the SSC docs from Stellar.org. In fact, this is actually very similar to a Distributed Trustless Worker Smart Contract proposed by Zulu Crypto.
In Zulu Crypto's smart contract, a customer is paying a worker for generating a vanity address (a computationally expensive problem). A very basic overview of how this works is:
- Customer creates an escrow account with 1000 XLM.
- Once the escrow account is funded, two transactions are created. The first transaction is a timebounded refund that will send the 1000 XLM back to the customer in case the exchange never takes place (this ensures the funds will not be lost). The second is a transaction that makes the escrow account escrowy aka the worker and customer are added as signers with weight 1, the thresholds of the escrow account are made weight 2 and the master key of the escrow is made weight 0.
- Worker proves they have found such a vanity keypair.
- The two create a transfer transaction where:
a. the worker sets the vanity account's signer to the customer, sets all the vanity's thresholds to 1 and sets the master key of the vanity account to 0
b. the user and the worker merge the escrow account to the worker so the worker gets paid.
Of course, a swap of physical goods would be a little more difficult. You would have to ensure that once the package is received, the payment is made. This may require some 3rd-party such as a delivery service, or a scannable barcode that activates the transfer transaction. Nevertheless, I imagine you could follow a similar pattern to the one outlined above. Maybe instead of a 2-of-2 multisignature, you could create a 2-of-3 where the 3rd party acts as an oracle.
What is an oracle? Here are some good reads:
tldr: I believe it is possible and a good starting ground may be to experiment with and build off of some of the already existing 2-party SSC's.