Zardnaar's analyses tend to be very focused on combat capability (not just whiteroom DPR, broader than that) rather than utility and flexibility, so definitely bear that in mind when reading them. I don't think that makes them valueless at all, but anyone's analysis is going to come from a specific place, and it's good to know what that place is.
A typical example might be some kind of infiltration into a fortress or castle in order to scout or take something or be in position to do something (like drop the drawbridge and sabotage the ).
A Rogue attempting this is going to, depending on what exactly is happening, and depending on the DM, be likely to be making an from a few to potentially dozens of checks, whether they're Stealth, Athletics, Deception (potentially with a bonus from the Disguise Kit though it's unclear if that really works in 2024, it'd be a cruel DM who didn't allow it), Persuasion, Slight of Hand, Thieves’ Tools, etc. etc.
Because of the way D&D works, every one of those is a binary pass/fail. Now it's often up to the DM to interpret that, but my long experience in D&D is that easily 90% of DMs read a fail as a total failure, like you fail at Deception, the guard doesn't just become suspicious or decide to come with you, they immediately draw on you and you're in combat. Perhaps that's a DMing issue but it's incredibly common one. Climbing and Stealth check fails can be particularly bad. A kind DM or one who doesn't want to make people rely on magic might allows you to get away with a very few checks, but again, a lot of DMs will be asking for multiple climb checks to climb a single wall (certainly if it's tall enough that you can't scale it in one round - and I've also seen a lot of DMs object when they find out exactly how fast PCs can technically climb in 5E), and god knows how many Stealth checks to sneak around.
Casters using spells like Fly, Spider Climb, Invisibility, Mage Hand, Transmute Rock and so on can often just skip to the end point of these plans, or skip like, 80% of the checks involved. There can be synergy, where spells are cast for the Rogue's benefit, but my experience is that you're lucky if the Rogue doesn't just feel like beneficiary of the caster.
Arcane Trickster isn't really an option, it doesn't offer much because its spells are so low-level, and imo suggesting it isn't very helpful, because the Rogue typically wants to be a Rogue, not a really bad Sorcerer about 1/3rd of their level. They do get invisible Mage Hand at least!
I feel like from how you're discussing this, you're looking at this from a "single-player" perspective, where you're in charge of all the characters, and those characters operate in perfect harmony and people always pursue the most resource-efficient solution (rather than the one that sounds like it's going to work best). But that's not how most D&D groups operate.
As for "not preparing and casting other spells", sure but that's rarely an issue. Sorcerers/Bards/Warlocks cause less of a problem because they have fixed spells and can't change them frequently, but if they did pick a spell that's good for this, they're more likely to want to use it.
