So, I've made a few small games in the past 3 years, and in particular have used external assets more than once, and had a few thoughts about doing it.
Be sure to follow any attribution requirements if you're using open assets that ask for it. At this point, I start a credits screen the moment I start pulling in external assets that have attribution requirements. You probably won't get sued by J. Random OpenGameArt poster, but it passes along goodwill, and also helps highlight their creativity.
1. Don't steal things you don't understand
To steal like an artist is to take something, and in some way, make it your own. That's a lot harder to do if you don't know any of the toolchain involved in editing whatever asset it is you're stealing.
So, if you're using pixel art, this means understanding which parts can or can't be changed without having to completely rework a given sprite. For 3D models, this means knowing how to work with the textures in question, or editing animations to help tune them for your gameplay. For audio/sfx, it means knowing your way around your engine's audio settings and/or Audacity enough to do things like remove clicking, or EQ things to fit into the soundscape of your game better.
2. Fit stolen things together
One of the defining qualities of "asset flips" is that they are clearly chimeras, constructed of things mashed together with no eye for consistency in visuals or work. So, if you're going to take assets from somewhere, make sure they fit together. Two obviously different pixel resolutions will stand out, as will mixing BFXR with foley off of OGA with no clear in-game reason for doing so. For 3D models, there may be some shader work that needs doing to help fit them into your art style.
3. Assets should gain you something
If you use assets, the time gained from not having to build them out should go towards making a better game, not just more game. If you're missing fundamentals in your game, for the genre it's in, no matter how pretty it is, or how cool that blood particle system is, a game that doesn't feel good to play is going to bounce a lot of players off (but see ). It's worth research into understanding common staples that make a given genre feel good. If you're going to make a Bad Game, you should at least know why it's a Bad Game
Here are some examples:
- Coyote Time
- Jump Buffering
- Clear understanding of jump distances
- Avoiding Jumps of Faith
- Collectables to encourage difficult routes
- Having a camera that keeps up with the player without jerking around too quickly
- Weighted Randomness to account for "feeling random"
- Good planning information for players (Slay the Spire and Into the Breach are great examples here, but even basic things like making unit roles clear are important)
- Clear attack visual feedback
- Readable scenarios
- Good character movement that isn't too slow (I'm not aware of a science for this, just that it needs up-front attention to get right, and that acceleration curves are part of it)
- Satisfying gun sounds
- Good enemy damage feedback
- Level design that communicates with the player well
- Avoiding symmetry in levels so the player doesn't get lost
Misc to avoid
- Upgrades that only change something by 1%, such that you can't tell the difference
- Uncertain player progress in the absence of narrative
4. Kenney (ne AssetJesus/Super Buddha) and other prior art
If in doubt, peruse Kenney.nl for assets. Kenney has more stuff than you'd think until you actually look, and I've used their stuff in at least 4 games that I can remember.
Also, this talk about The First Tree is well worth a watch, and discusses the use of assets at some length (and makes some of the same points I do)
5. Games are Ugly until the End Anyway
Don't be afraid to hold off on looking for assets for a game. Get some placeholders in, and know where you want certain things to go, but unless you're in a jam, don't sweat the asset quality until you've found the fun, as it were.
 : Sometimes you want to make a game that isn't super smooth/fun for a number of reasons, like making a horror game, or trying to communicate a frustrating experience to the player. But, the more you do that, the more invested the player has to be to stick around. Pathologic stands out as an example of a game that does a lot of things that are outside of common wisdom, but that has a dedicated following singing it's praises. Dark Souls can take a long time to get used to, but rewards player investment in a lot of ways as well.