The gemara says that Rav davened ma'ariv for shabbos on erev shabbos. However, the gemara justifies this by using the opinion of R. Yehuda that after plag hamincha it is already considered to be night. We only find that R. Yehuda says this in regard to tefillah, not for anything else. Tosafos and the Rosh cite Rabbeinu Tam (beginning of masechta) use this to justify saying shema early as well. However, the Rosh takes issue with this Rabbeinu Tam that for krias shema, plag hamincha won't help. In short, it is only clear that one can rely on this leniency of davening early on erev shabbos for tefillah, not for anything else. It is clear from the gemara that there would also be an issur melacha (unless the kabala was by mistake), but it isn't clear if the issur melacha would be d'oraysa. Rabbeinu Yona seems to be mesu'pak whether the issur melacha during the time of tosefes would be d'oraysa.
But, the gemara then goes on to day that Rav even made kiddush early on Friday night. Kiddush is a mitzvah d'oraysa that must be done on shabbos. How is it possible to make kiddush early? The fact that even kiddush can be made early (shulchan aruch 267:2) implies that one can actually transform the day into shabbos by being me'kabel shabbos early. Although the M.B. (5) cites two opinions as to whether one should make sure to eat some of the seuda when it is actually shabbos, that is because the pasuk that is used as the source for the 3 seudos uses the term "yom" - day. But for kiddush it seems that one can certainly be yotzei during the z'man of tosefes. It would seem that even if we assume that Tosefes shabbos is only d'rabonon regarding the issur melacha, as Rabbeinu Yona suggests, the z'man would still have kedushas shabbos on a Torah level to enable you to fulfill the mitzvah of kiddush.
Rav Moshe (Igros Moshe O.C. 3:38 - end of teshuva) raises a very interesting point. There is a well known halacha (O.C. 263:12) that when there is only one shul in town that is mekabel shabbos early, the entire community must begin shabbos with the tzibbur and one cannot continue doing melacha after that time. Rav Moshe says that he isn't sure if this din applies nowadays when the shul is mekabel shabbos early only for convenience. He suggests that perhaps the din of the tzibbur being pulled after the shul being mekabel shabbos is only when they are intending to fulfill a mitzvah of tosefes shabbos, not when it is purely for convenience. Therefore, when it is only done in the summer and not the winter, it is a clear indication that the intent is not for the mitzvah and may not bind the entire community to an early shabbos.
It has been troubling me where Rav Moshe gets this notion from. Why would he think to make a distinction based on the intent of the shul being mekabel shabbos? I found that Rabbeinu Yona in explaining Rav davening early assumes that he would say shemoneh esrei early, but say shema with it's brachos after tzeis hakochavim. He asks, what about סמיכת גאולה לתפלה? To which he answers that since Rav was intending to fulfill the mitzvah of tosefes shabbos, he was willing to forfeit the advantage of being סומך גאולה לתפלה. Now the question is, how did Rabbeinu Yona know that Rav was intending to fulfill the mitzvah of tosefes shabbos, perhaps he too was doing it only out of convenience? The gemara says that Rav had a practice of davening early from which we see he held like Rav Yehuda, but Rav Huna didn't have this practice so we see he doesn't hold like Rav Yehuda. The Tzlach asks that we don't see anything from Rav Huna. Perhaps Rav Huna also agrees that one can be mekabel shabbos and daven after plag hamincha but didn't do this because he didn't want to! The Tzlach answers that since there is a mitzvah of tosefes shabbos, the assumption of the gemara is that if Rav Huna held it can be done, he would have done so. We see based on the approach of the tzlach that the gemara itself assumes that the compelling reason to be מצלי של שבת בערב שבת was to fulfill the mitzvah of tosefes shabbos. This is perhaps why Rabbeinu Yona assumes that Rav "intended" to fulfill the mitzvah of tosefes shabbos.
Based on this we can explain Rav Moshe's suggestion. The source of everyone in the community being drawn after the shul is found in the Mordechai (Shabbos perek 2, 297). Since the only precedent we find in the gemara for accepting shabbos early is when one is intending for the mitzvah, it is fair to assume that only in that case the Mordechai citing the Riva"m would consider the individuals in the community to be drawn after the tzibbur.
Another interesting point is that the Mordechai is assuming that the majority of the community were in shul, just that there were some individuals not in shul. Perhaps in a situation where there were more observant Jews who weren't yet mekabel shabbos, not in shul, they would not be bound to the kabalas shabbos of the shul.