An attacker aiming to lower the feerate requires low-feerate ancestors; however, these are prone to being dropped from various network mempools. This dropping leads to the possibility of an honest transaction with a standard feerate proceeding unobstructed, as it would not be considered a double-spend in those mempools. This observation is supported by @TheBlueMatt's practical experiments, which suggest that even transactions with relatively low feerates can propagate and eventually be confirmed.

Another point of concern is witness inflation, which targets the joint transaction without depending on unconfirmed ancestors. The vulnerability persists in that an attacker could inflate the replace-by-fee (RBF) cost for a victim by creating a single 1000 virtual byte (vB) descendant transaction. By repeating this process, the attacker could potentially deplete the victim's funds through fees without incurring any costs themselves. Although it is uncertain whether this method can be consistently effective, the only way to confirm its viability is through attempting such an attack.

While version 3 (v3) of the protocol does offer protection against pinning, it appears insufficient in preventing witness inflation attacks, indicating that there is still room for improvement in securing Bitcoin transactions against these exploitation tactics.