I went a bit further looking for a foreign blockchain that would allow a proof of concept.
I realized how well is multisignature implemented in Stellar. The design I proposed require on both chains a native multisignature feature that allow to add/remove signers over time to prevent anchor decay.
I didn't found that elsewhere. So I guess a proof of concept would have to be implemented between Stellar public and testnet. At this point the trustless anchor design could at least provide migration of native assets between Stellar networks, which may be of some use later if some other organizations happen to implement their blockchains using Stellar.
There's also the possibility to implement multisignature or an alternative solution using smart contracts. This is possible. The drawback here is that smart contracts are tricky and the possibility of a failure (i.e.: all money blocked or gone) is non-null, which is not acceptable in a trustless design. Ethereum gave us an example of that with parity wallet failure.
Hopefully someone will come with some clue to go forward with this.