As Charanor and Philipp point out in other the comments & answers, there is a school of thought in game design (called "Love the Player" in my studio) that says if the player wants to do something that doesn't break the game for other players, err on the side of letting them do it. Players who see a kidnapping twist coming and strategically prepare for it get to feel smart, rather than powerless, and you avoid creating a situation where the player's one-of-a-kind ultra-rare gear can be lost forever without recourse, without needing to create extra systems to protect against this. That's a valid approach to this design problem.
There's another school of thought, articulated in Soren Johnson's article "Water Finds a Crack," that says that players are drawn to choices that give them a material benefit in games, even to the detriment of their own enjoyment. Players tend to undervalue their subjective experience and time investment relative to things they can quantify in a game, like the value of gear. This means that some players who would normally leave the gear on - for reasons of role-playing immersion, or just to avoid extra menu fiddling - can feel pressured to play differently because that's the "wrong" choice according to the numbers.
If this proves to be a concern for your target audience (validate this by polling testers on their subjective impressions - if they're engaging in gameplay you'd consider "wrong," but enjoying it all the same, then maybe it's not a problem for them), there are things we can do to address this dissonance.
The other answers do a great job covering several different approaches, so I want to suggest just one more: narrow the gap between the player's reality and the character's.
If this were a real life group of medieval adventurers, one of the
group members wouldn't be ordered to strip down to their underwear
moments before being ambushed and captured.
Why not? What consequences lead the characters to make different decisions than players in this case? And how can we reintroduce those consequences into the player's view of the situation? Some thoughts:
Morality: taking all your ally's stuff and leaving them to rot is a pretty dick move. It could sit poorly with a character's conscience or moral code (alignment system) or affect how likely other people are to trust them (reputation system) or possibly even invite retribution from the wronged party and their allies (revenge system)
Uncertainty: the characters don't know when the ambush will happen, so they can't plan for it. Taking away useful gear on just the possibility that a character might get abducted means their utility and survivability in fights leading up to that event are greatly reduced. Set the abduction scene to happen sometime randomly within a gauntlet of several encounters with no save point in-between, and you put even a player with access to walkthroughs in a similar position. It's still possible to savescum it, but the increased difficulty/time investment can help players who aren't really there for the hyper-optimization to resist the temptation.
Future Payoff: If I leave a companion with good gear, I can hope they'll survive long enough to meet & help me again. Or for them (or their next of kin) to repay my generosity in other ways. Make the gear the character left with matter in some way - maybe you play a short vignette mission as that character while they're separated from the group, so stripping them down before that actually makes this harder to proceed (just beware of trapping a hasty player in an unwinnable situation if they can't backtrack/reload to before the split). Or maybe you model the chance the character survives, or their prosperity while they're away from the group, based on a function of the gear they were left with. A sidekick let loose with good gear is more successful in developing their skills and is higher-level / has even better gear when you next meet them, for example.
