This is the sort of thing that drives home some marked differences in assumptions, systems and play approaches.Player: My uncle told me (player consults the module the DM is running) that if we go down the left corridor and open the second door, under the rug is a secret compartment with 85gp and 2 potions of healing inside.
You don't see anything wrong with that? Weak justifications for metagaming are just that. Weak justifications. There absolutely does need to be something further in order for the character to have that knowledge.
In particular:
(1) To me this does not register as an issue of metagaming but (at least as I understand what is being presented) as cheating. In running a classic module of the KotB or Castle Amber or Desert of Desolation sort, it is understood that the player does not read the GM-only material, because this is the puzzle the player is expected to solve. Doing otherwise is cheating.
(2) This presupposes that the location of some treasure in a geographical location is established by the GM in advance, but not announced to the players, such that the player might take steps (like peeking at notes or reading the module) to learn it. In a game like Cortex+ Heroic RP, and I would imagine in many DW games, that presupposition does not hold good. In Cortex+ Heroic, for instance, most instances of treasure are going to be either assets or similar established by players as part of the action resolution process, or Scene Distinctions whose existence is clearly announced to the players as part of scene framing.
(3) This presupposes that a player is free to decide what it is that his/her uncle has told the PC, and its usefulness. But in (say) DW, this is not the case: this would be a Spout Lore action and so requires a check as discussed at some length upthread. In BW, the GM could say yes but equally could call for a check (eg on My old uncle's stories-wise) which, if it failed, would license the GM to have some fun with the player about the tall tales told the PC by his/her uncle!
(2) This presupposes that the location of some treasure in a geographical location is established by the GM in advance, but not announced to the players, such that the player might take steps (like peeking at notes or reading the module) to learn it. In a game like Cortex+ Heroic RP, and I would imagine in many DW games, that presupposition does not hold good. In Cortex+ Heroic, for instance, most instances of treasure are going to be either assets or similar established by players as part of the action resolution process, or Scene Distinctions whose existence is clearly announced to the players as part of scene framing.
(3) This presupposes that a player is free to decide what it is that his/her uncle has told the PC, and its usefulness. But in (say) DW, this is not the case: this would be a Spout Lore action and so requires a check as discussed at some length upthread. In BW, the GM could say yes but equally could call for a check (eg on My old uncle's stories-wise) which, if it failed, would license the GM to have some fun with the player about the tall tales told the PC by his/her uncle!
Thus, this example is located in a very particular play paradigm. There are a variety of other systems and approaches to which it does not straightforwardly generalise.
This also relates to the discussion, upthread, of whether or not D&D is "strong GM decides". To the extent that it permits this sort of thing to come about, it certainly lacks a whole suite of action declaration and resolution systems that other systems use to manage action declarations about stuff my uncle told me.