I would also like to know how TAU is protected against the following hypothetical attacks:

Greedy Flock Problem

I notice you have now changed the whitepaper to say that harvest power is now determined by the cumulative transaction fee paid in a pre-defined window. However, in the early days of the TAU network, this can easily be exploited by a determined individual to gain majority control over the block rewards. For example, if there a total of 10 harvest clubs existing of equal size (in terms of harvest power per unit time), then we would expect the rewards distributed to each of these clubs to be approximately equal. However, if one of these clubs suddenly offers a large incentive for users to switch (the early bird), then this will lead to a massive reduction in number of harvest clubs as everybody flocks to the single incentivized club. This would essentially easily allow one relatively high net worth individual to take over the entire network by incentivizing harvest club switching.

False Incentives

You mentioned previously that some harvest clubs would profit more on reputation rather than the transaction fees, which might actually lead them to incentivize club members by taking a loss. However, this leads them open to being sniped by other malicious groups which offer false incentives to get new members, e.g. $100 BTC per member, then chooses to ignore signal transactions to lock them in. If properly designed, this scam could lead to the creation of high harvest power clubs which members cannot escape from easily, whilst the club leader takes the entire reward.

Potential Resolution: Terms of the harvest club should be displayed to club members via a signal transaction from the club leader. This signal transaction would include data pertaining to the harvest reward schedule offered by the club.

The Bait and Switch

Under the new cumulative fee window system, there is an additional possible attack where a harvest club leader can achieve almost entire monopoly over the harvest rewards. Imagine a scenario where a successful harvest club manages to achieve among the highest cumulative fee, therefore achieving a sizeable fraction of future rewards. The club leader can then simply claim that the harvest club is shutting down, leading everybody to leave the club to join another. However, this was simply a bait, once all the club members have left, the club leader will then be in full control over the club, and hence be the sole beneficiary of the block rewards.

Potential Resolution: Implement a voting system to force change the harvest club leader.

Overall, I feel like the signal transaction concept is under utilized. Signal transactions should also be used to carry information relating to the parameters of the harvest club to its members and notify them of any changes. It would also be a good idea to allow people to set a timer on their harvest power delegation, either by automatically sending another signal transaction on X data (via the wallet), or by including a self-destruct parameter in the signal itself.

Identification vulnerability
I suspect that the signal transaction feature, though quite useful for the purposes of TAU, will also enable global blockchain analysis companies to easily track and identify relationships between wallet holders, as signal transactions can be considered a more intimate transaction than a standard one, therefore highlighting friendships, business relationships and partnerships, rather than commerce transactions. I suspect public visibility of signal transactions on the blockchain will eventually lead to a much more complete database of ID/Wallet holders compared to other similar projects such as Bitcoin. I think this can be resolved by ommitting one-half of the wallet details on the blockchain (or possibly both halves).

Also, there are issues whereby the harvest club leader can force its club members to collude with it for nefarious purposes by holding the block rewards hostage. For this reason I think a voting system is definitely necessary.