Contract-level Relative Timelocks (or, let's talk about ancestry proofs and singletons)

The discussion raises a concern regarding the vulnerabilities in transaction management systems, particularly focusing on the potential for malicious actors to manipulate transaction states to their advantage. When Mallory, representing the adversary, submits an older state of a transaction, and Alice counters by submitting the latest state from the mempool, Mallory can exploit the situation by outbidding Alice's submission with any previously submitted states. This scenario underscores a critical flaw in systems that rely on TXID-dependent fee-paying mechanisms. In such systems, Alice is compelled to re-sign her transactions each time Mallory outbids her, despite her signatures remaining valid. The transactions themselves are removed from the mempool, indicating a failure in the system's ability to securely manage transaction states against adversaries.

The critique extends to the effectiveness of Commitment-Ledger-Roll-Tree (CLRT) mechanisms in addressing these vulnerabilities. Contrary to expectations that CLRT would offer mitigation against such attack vectors, the observation suggests that it may inadvertently exacerbate the risk. By allowing previous states to be used as leverage in bidding wars over transaction validity, systems employing CLRT could be providing malicious actors with an even more potent tool for undermining transaction integrity. This analysis reveals a pressing need for revising current approaches to transaction management and fee payment in blockchain systems to safeguard against such exploitation techniques.