A few years ago, Phil Manning at the University of Manchester used a robotic simulation of the dromaeosaur sickle claw to determine how it may have been used - he wanted to test the assumption, popularized by Jurassic Park, that it was a disemboweling tool, perfect for slashing through the bellies of prey dinosaurs. The robo-claw wasn't actually able to slash through the flesh of a pig's carcass; rather, it punctured it. This led to the conclusion that raptors most likely used the claws to maintain a grip on their larger prey while ripping them up with their teeth, leading to a pretty awful, painful death by bleeding.
Manning is continuing to explore the use of dromaeosaurs' claws, and a recent study of the claws of their hands, published in the newest, dino-centric issue of the Anatomical Record, suggests that they could have supported their weight when climbing. In particular, the paper discusses the manual (hand) claws of Velociraptor mongoliensis, whose inner curve had an arc of 127 degrees, which falls between the upper limits of perching bird claws and the lower limits of trunk-climbing bird's claws. Manning also refers back to a 1969 study of Deinonychus antirrhopus, another small dromaeosaur; their pedal (foot) claws had an arc of 160 degrees, which is in line with the upper range of today's trunk-climbing birds, like woodpeckers or the nuthatches which frequent my suet block. Further studies will be devoted to Deinonychus foot claws, of which we have more good specimens than those of Velociraptor.
So the dromaeosaurs, at least the smaller ones, may very well have used their claws to climb trees as well as hang on to their dinners. Lacking any proper paleoart depicting this fascinating new idea, here's an incredibly crappy, dashed-off, proportionally inaccurate one from yours truly.
Image by me, via flickr