Mass: We can only measure mass directly for stars in binaries, and in many of those cases we suspect the companion influences phenomena seen on the primary. alpha Cen is an exception where we have two well-separated stars with highly accurate masses (Kamper & Wesselink 1978).
Composition: The elements that most affect stellar structure are C, N, O, and He, but these are the hardest to measure. At present we know best the heavy-element abundances (Fe, Ti, Mg, Ca, etc.), but that index can be determined to about 0.05 dex, as indicated by internal consistency for cluster stars (Boesgaard & Friel 1990).
Age: Only for the Sun do we know age in a fundamental sense because we can apply knowledge of the decay of radioactivity to meteoritic samples. The next best case is probably alpha Centauri because we know the masses of the stars and their abundances and they are similar to the Sun. Even for alpha Cen derived ages disagree by 1 Gyr or more (i.e., about 20%) (Furenlid & Meylan 1990; Noels et al. 1991; Neuforge 1993). We are reasonably confident of the ages of open clusters like the Hyades or Pleiades, but even there different computations of the physics of the high-mass stars that are the determinants of the cluster's age can produce differences in age of 20 to 25%.