Look for something in the category of Humanist Sans Serif typefaces. These will tend to have more open apertures, making it easier to distinguish numerals 3 6 8 9 from each other. If budget is a consideration, there are a number of libre (open source) fonts that you could try, e.g. Noto Sans.
I do design typefaces too. I'm not saying it's not important to protect the thing we do. I have seen a couple of times that the exact unique thing one of my typefaces had had been distorted.
Oh, I've seen my Gabriola typeface fake bold more times that I've seen it displayed as intended; also obliqued, stretched, squished, rendered in neon — okay, so that was pretty cool! —, and amateurishly handpainted.
I think the assumption that concerns about quality are concerns — or primarily concerns — with reputation are mistaken. It isn't that I think that someone will see a poorly modified font and associate it with me, but that I don't like to see things created with care and attention to detail being messed up. For me, the concern about quality is concern about the thing itself.
This is not to suggest, of course, that all modifications result in a diminishment of quality, but it does mean that if one has a blanket permission for modification in a license, then one is surrendering any kind of quality control. It's extra work, but I think requiring a licensee to contact you and ask for permission to modify a font is a reasonable requirement. It allows one to ask some questions, offer some advice, or even propose to make the modifications oneself.
re. co-ownership of derivative works:
Couldn't the EULA cover that?
No, I don't think so. Or, rather, I don't think any lawyer would advise that. The EULA is a license for use agreement, not a copyright assignment agreement. It would be very unwise to include in the license agreement any kind of upfront copyright assignment for derivative works that have not yet been created, without any knowledge of the nature of those derivative works, who will be making them, under what circumstances, in ways that put restrictions on one's own rights.
Unless a license explicitly prohibits analysing the kerning data, where's the infringement? Collecting and analysing kerning data doesn't even necessarily involve decompiling the font, since it can be collected from text output. Kern values, as implemented in the kern table and typical GPOS pair adjustment lookups, are data, which typically has minimal copyright protection. In the US, for instance, I believe it is still the case that data itself is not protected by copyright, only particular selections or arrangements of data. So a kern table might be protected as part of the font software, but the kern values in that table would not be. [Usual caveat: I am not a lawyer.]