Now, I'm a fan of universal mechanics and feel they're one thing from latter permutations of DnD to import into old school play. However, a universal mechanic has some pitfalls. I say this as a DM that has played those latter permutations and is now running a game that harkens back to old school ways. I try to challenge the players, not just the characters.
As a DM, I like the convenience and freedom of deciding how difficult a task is and having a framework to place that in. The essential elements of the framework are a scale of difficulties (including opposed checks) and measurement of character ability. During play this means that I can concentrate on creating, listening and adjudicating without pausing the game to look up a rule. The players also don't need to be confused unduly and have some expectation of how their characters function. This is one of the reasons that I reached for Castles and Crusades when I began co-DMing our Greyhawk/Yggsburgh/Castle Zagyg/Mad Archmage campaign (even though I'm house ruling SIEGE more and more).
A universal mechanic makes it easier to allow players to try crazy things - a joy that makes the game.
But there can be too much of a good thing. Before calling out target numbers and telling players to roll, ask yourself following questions:
1. Is this something that adventurers would fail at?
While we in the old school are often known for our love of the grim and gritty, adventurers are still decently competent at their jobs. "Giants in the Playground" pokes great fun at the absurdity of spot and listen checks in DnD 3.x. Now, even OD&D had spot checks of a sort - they're hidden in rolls to find secret doors and in the rolls for surprise. Still, with the ease of the mechanic the various "perception" checks are definitely over used.
I've seen by the book old school over use mechanics as well. Is EVERY door really that stuck? Likewise some supposedly old school DMs would rather have the players roll a d6 then describe the book shelf that hides the wizard's secret door.
Also, most old school games are based around classes. To be useful and meaningful, classes imply a wide set of skills in which the character is competent. A thief is a stealthy individual and not every attempt at moving silently should be subject to the fickle whims of a d20 or a percentile roll. Likewise a fighter should be competent about combat and weapons, a magic-user knowledgeable about arcane lore and what not.
2. Is this challenge more interesting than a number?
Matt Finch wrote the book on this one, though I don't think he stated it in this fashion. We game in fantastic locations and dream up personalities and magic and mystery. Aren't they more than mere numbers?
In fiction, secret doors are fascinating, fun conundrums. Finding the secret door to Batman's crime lab is a part of the game. The traps in James Raggi's "The Grinding Gear" are fiendish, clever and fun. It would be a disservice to your players if they could bypass them with a roll of a d20 instead of interacting with that strange, weird and dangerous place.
The Tomb of Horrors is more than a set of absurdly high numbers to roll.
3. What does success mean? What does failure mean?
If you allow your players to roll, you need to be prepared to tell them something cool happens when they roll high and beat the difficulty number in your head. Never allow the player to roll for something you think is not possible, because fate will give you a "20" to adjudicate.
Similarly, if you can't think of an interesting or meaningful failure result you should just let the character succeed.
Lastly: Is rolling for this how I want to spend my time?