From Ephraim Stulberg:
(ל"א:ט"ו)
שֵׁשֶׁת יָמִים יֵעָשֶׂה מְלָאכָה
In the above פסוק, where we are told that "for six days shall work be performed." The passive "יֵעָשֶׂה" is of course masculine, in spite of the fact that its subject is the feminine "מְלָאכָה". And so follows the question: What's up with that?
R, Yaakov Kamenetsky observes that in many instances in which the word "כל", or "all", is used to modify a noun, the gender of the verb/adjective used to describe even a feminine noun will be masculine. Thus, in פרשת בא, we have the phrase (כָּל מְלָאכָה לא יֵעָשֶׂה בָהֶם" (י"ב:ט"ז", where the masculine "יֵעָשֶׂה" refers back to the word "כל" more than it does to the feminine "מְלָאכָה". R' Kamenetsky suggests that the verse in כי תשא is to be read as though the word "כל" were present. R' Kamenetsky also quotes the opinion of a certain R' Nathan, who appears to have been some sort of confederate of his back in Lithuania, who explains the incidence in בא by positing a rule in which the gender of passively constructed verbs does not necessarily correspond to the gender of their related nouns. R' Kamenetsky is somewhat dismissive of this suggestion, though it clearly solves the question in כי תשא much more effectively.The truth is that R' Nathan's suggestion had already been anticipated by an earlier authority, namely רד"ק, in his comments on מלכים א ב:כ"א:
וַתּאמֶר יֻתַּן אֶת אֲבִישַׁג הַשֻּׁנַמִּית לַאֲדנִיָּהוּ אָחִיךָ לְאִשָּׁה
He explains that when the passive voice is employed, it creates a sort of gap between noun and verb. רד"ק reads the verse in מלכים as follows: "It shall be given, namely Avishag the Shunamite, to Adoniyahu". Likewise, we would read: "For six days it shall be done, namely 'work'". Actually, this is not very different from the explanation given by R' Kamenetsky. R' Kamenetsky makes the important point of noting that the phenomenon is not limited to the נפעל, and that it really applies in any case where a neuter noun is employed. However, R' Nathan's point is also crucial, for it recognizes that this phenomenon will be much more prevalent in cases in which the נפעל is utilized, thus creating an implicit break between subject and verb which is filled by the invisible neuter.