1) In some cases the TAF/METAR symbols next to an airfield on the map shows "M" even though the weather is better than the particular "Good VFR" settings. Example: Assume "Good VFR" minima settings to be OVC/BKN and SCT both 2000 feet, VIS 8000 m. Then the following TAFs/METARs all give an "M" symbol.
ESMX 150230Z 1504/1512 18010KT 9999 -RA BKN035
ESSL 150450Z 18007KT 150V210 9999 RA FEW049 BKN062 06/04 Q1020 REDZ
ESCF 150450Z 18007KT 9999 -RA NSC 07/05 Q1019
It guess that the observed/forecast rain automatically gives "M"? This doesn't seem very useful. I don't consider light rain with VIS 10 km or more and "no significant cloud" to be marginal...
2) It appears that Skydemon ignore change groups in TAFs. E.g. everything after the first BECMG, TEMPO etc. is ignored. I would expect the program to use the given departure time (if any) or the time set in the "Weather" window.
3) TAFs issued by MIL MET offices in Sweden include forecast QFE and QNH figures, e.g.
ESDF 170830Z 1709/1718 25005KT CAVOK QFE 1011 QNH 1018
I certainly don't expect Skydemon to understand such non-standard format, but it would be good if it could be more defensive in interpreting non-standard TAFs. In this case the four-figure QFE value (1011) is taken to be the horizontal visibility. (Both the TAF decode and the weather symbol on the map shows this.) As CAVOK implies at least 10 km visibility, there can't be a separate visibilty figure, so Skydemon should ignore what looks like a visibility figure following "CAVOK" (just as it appears to ignore the "1018" figure).
4) Skydemon doesn't understand the SNOWTAM part of Swedish METARs. E.g. the METAR
ESSA 150450Z 18015KT CAVOK 06/02 Q1020 R88/19//95 NOSIG
gives an error message "Unrecognised part "R88/19//95" included in METAR ". (This is certainly a very minor complaint.)