Black has played …a6 and …e5 in a Sicilian, White’s knight has stepped back to b3, and d5 is now a hole. Look at which black pawns could ever attack that square: the c-pawn, traded off on move three, and the e-pawn, which has just marched past it. Neither is coming back. Pawns do not reverse.
8 
7 
2 
a c 
d 
e h 1 That is the definition, and it is a purely mechanical test. A square is weak when no pawn of yours can ever be brought to attack it. Not “does not attack it now”. Ever.
Where holes come from
Every pawn move you make gives up two squares for good. Push the e-pawn from e2 to e4 and d3 and f3 are no longer yours to defend with it; play …g6 and f6 and h6 turn permanently dark. Most holes in club games come from three habits: pushing a pawn in front of the castled king, fianchettoing and then trading the bishop that covers the same colour, and making a capture that leaves a whole file empty.
The Sicilian hole above is a deliberate one. Black accepts d5 because he gets the half-open c-file and fast play in return, and thousands of strong players make that trade happily. A hole you have thought about is a decision. A hole you never noticed is a weakness.
What lives in one
A knight, almost always. Put a white knight on d5 in that position and it hits b6, c7, e7, f6, f4 and e3 while nothing can chase it, and Black spends the rest of the game arranging his pieces around it.
7 2 a c d f g That square with a piece parked on it is what the other side calls an outpost. Same square, different chair.
Before you push a pawn
One question, every time: which two squares does this pawn stop guarding, and can anything of his get to them? If the answer is a knight, in three moves, then think again or make sure you get something concrete for it. That habit is the cheapest positional skill there is, and it will cost you less time than a single middlegame puzzle.