Submitted by t0x1c, also found by 0xpiken, klau5, and haxatron
https://github.com/code-423n4/2024-01-salty/blob/main/src/stable/StableConfig.sol#L51-L54 
changeMinimumCollateralRatioPercent() allows the owner to change the minimumCollateralRatioPercent while the changeRewardPercentForCallingLiquidation() function allows to change the rewardPercentForCallingLiquidation. 
The protocol aims to maintain a remaining ratio of 105% as is evident by the comments in these 2 functions:
and
That is, if borrow amount is 100, after the calls to any of these two functions, there should be at least 105 collateral remaining which will act as buffer against market volatility.
The current calculation however is incorrect and there is really no direct relationship (in the way the developer assumes) between the rewardPercentForCallingLiquidation and minimumCollateralRatioPercent. Consider this:
A table outlining the real buffer upper limit values is provided below. Another table showing the actual desirable gap in values is also provided so that the buffer always is above 105%. 
Straightaway, it can be seen that the current default protocol values of 5% and 110% give a buffer of less than 105% and hence either the minimumCollateralRatioPercent needs to have a lower limit of 111 instead of 110, or there should be agreement to the fact that 103.5% is an acceptable remainingRatio figure under the current scheme of things.
All figures expressed as % of borrowed amount i.e. 104.5 means 100 was borrowed. 
Current Implementation:
Desired figures for maintaining an upper limit of >= 105% of borrowed amount:
Assuming that the protocol wants to calculate an actual 105% remainingRatio, changes along these lines need to be made. Please note that you may have to additionally make sure rounding errors & precision loss do not creep in. These suggestions point towards a general direction:
Update the two functions:
and
Also L39:
Picodes (Judge) commented:
othernet-global (Salty.IO) acknowledged and commented:
Status: Mitigation confirmed. Full details in reports from zzebra83, 0xpiken, and t0x1c.
