There are quite a variety of concerns here. For a full overview you can check "Lorentzian wormholes" by Visser. For now we'll consider wormholes defined by the thin-shell formalism, where the matter propping up the wormhole is on a very thin shell on the wormhole mouth.
First, there's the tidal forces. As the tidal stress induced by gravitational forces depends on the Riemann tensor, we can consider the Riemann tensor of a thin-shell spacetime :
$$R_{abcd} = - \delta(r) \tau_{abcd} + \Theta(r) R^+_{abcd} + \Theta(-r) R^-_{abcd}$$
with $\tau_{abcd}$ a tensor depending on the discontinuity of the extrinsic curvature $K$. Since this is the most important part in propping the wormhole open, let's assume (without any good reasons) that we can neglect the curvature outside of the wormhole mouth.
The extrinsic curvature tensor locally looks roughly like
\begin{equation}
K = \begin{pmatrix}
1/R_t & 0 & 0 & 0 \\
0 & 1/R_1 & 0 & 0 \\
0 & 0 & 1/R_2 & 0 \\
0 & 0 & 0 & 0 & 0
\end{pmatrix}
\end{equation}
$R_t$ is the radius of the timelike curve described by that point (it's related to the wormhole's acceleration), while $R_1$, $R_2$ are the principal radii of the surface. Hence by making the radii as large as possible (that is, making the surface as flat as possible), we can make the tidal effects arbitrarily small.
In other words, if the wormhole mouth has at least one part of it that is flat, then one can pass through it unmolested by tidal forces.
Then there's the matter of the matter itself propping the wormhole open. As there isn't really any realistic model of matter that can prop up a people-sized wormhole (or even one large enough to get one electron to go through), it's hard to say exactly what are the possible risks (the matter involved might not even couple to ordinary matter), but considering a rough estimate of the order of magnitude for the (absolute) energy is about $10^{27} kg$, in an area of a few cubic meter, it's safe to say that it's probably not a great idea to approach it too much.
There again, flat faces are a benefit : the stress-energy tensor for the thin-shell part is again zero if the surface is flat. This does not mean still that this is a good idea : such very dense matter is likely to radiate in the vicinity.
Another issue is the throat length and time : if we drop the thin-shell assumption, which isn't terribly realistic, there is still some way to travel between the two mouthes. The length of the throat is related to the amount of negative energy required (long throats require less negative energy), which means that a more reasonable worhole may have very long travel times to actually cross. The perceived time is also an issue, as the travel time could be very short for the person crossing it but very long from the outside perspective (this is related to the redshift function which brings up a whole bunch of other potential problems).
Then if we actually assume a time machine scenario, things get much much worse. To simplify matters somewhat, you might be aware that you can describe the quantum vacuum as a gallore of virtual particles (wrong yes, but let's go with that for now). As a wormhole approaches time machine formation, there are more and more virtual particles forming almost-loops, which has the bad side effect of blue-shifting them, increasing their energy. Those become actual loops when the time machine is actually formed, which has the result of making the stress-energy tensor diverge : the quantum vacuum has infinite energy, which probably bad results if this was possible.
It's quite likely that the wormhole collapses before this happens.