Enable qKSM and qUSDT as collateral for Kintsugi Vaults

What will happen if the vault going to be liquidated, but all his qAseests are actually borrowed by another Kintsugi users?

Ex:

The guy have 3 BTC in vault and 135K qUSDT collateral.
Imagine the BTC rised 10% from 31K atm to 34K.

Then that mean the 135% threshold done and the guy should be liquidated.
However lets say all his 135K USDT are borrowed by some others.

All that prolly mean the liquidator would liquidate to get an asset he actually can't move/change or smth.

One more question is about volatility impact. Right now we have 20K liquidiy in BTC/USDT pool, wouldn't 10K USDT sell-off drop USDT price to the threshold where some BTC vaults with USDT collateral can get liquidated?

What will happen if the vault going to be liquidated, but all his qAseests are actually borrowed by another Kintsugi users?
In the pooled lending model (implemented), there is no 1:1 mapping between borrowers and lenders as in P2P lending. This means: anyone can withdraw on a first-come-first-served basis if there is free capacity. If there is no capacity left this means utilization is at 100% and the interest rate will be very high: lenders earn high interest rates while borrowers are incentivized to repay their loans to avoid high fees.

The risk of this happening can be further reduced by setting the target utilization rate lower (e.g. 80 or 70% instead of 90%).

One more question is about volatility impact. Right now we have 20K liquidiy in BTC/USDT pool, wouldn't 10K USDT sell-off drop USDT price to the threshold where some BTC vaults with USDT collateral can get liquidated?
The price feed used for liquidations comes from oracles (e.g. DIA), so manipulating the Kintsugi dex will have no impact on Vault liquidations. In the future, as liquidity becomes deeper, it may be possible to use the DEX as additional price feed - but right this would be too risky as you correctly point out.

Put my vote however ^^