Possible evolution of feather distribution and feather types based on current knowledge, counter clockwise from bottom left (not to scale):
Whatever-Juravenator-and-Sciurumimus-are grade (represented by Juravenator starki) - Scales on tail and hindlimbs. Protofeathers or plumulaceous feathers on at least the top edge of the tail and probably much of the body. (A lot of the fossilized feathers formerly assumed to be "protofeathers" are probably more advanced than they look. Studies on the effect of crushing on bird carcasses show that, when crushed, the pennaceous body feathers of modern bird resemble the supposed "protofeathers" of basal coelurosaurs. So the "protofeathers" in basal coelurosaurs are likely at least plumulaceous feathers in reality, while those reported in more derived taxa are almost certainly true pennaceous feathers.)
I know, many suggest that feathers originated at the base of Dinosauria due to the presence of feather-like integument in certain ornithischians, but even if the fuzz of heterodontosaurids and theropods are homologous, we do not know enough about the feathers (if present) of taxa outside of whatever the least inclusive clade including both Juravenator/Sciurumimus and maniraptors is (possibly Coelurosauria, possibly Orionides if Sciurumimus and Juravenator are megalosauroids) to say much on the subject. (There's another non-coelurosaur theropod that's said to preserve evidence of feathering on the arms in the form of quill knobs, Concavenator corcovatus, but this is dubious. The "quill knobs" are on the "wrong" side of the ulna, are irregularly spaced, and appear to match an intermuscular line that's present in crocodilians, so they might be for muscle attachment instead.)
Compsognathid (and lots-of-basal-coelurosaurs-that-might-be-compsognathids) grade (represented by Sinocalliopteryx gigas) - Protofeathers or plumulaceous feathers on tail, torso, forelimbs, and legs, extending onto the foot in at least Sinocalliopteryx. Integument on underparts and snout unknown. Like Juravenator, Compsognathus longipes had scales on at least part of its tail, but this is not known in other taxa from this grade. Maybe they had scales on the sides and underside of the tail with feathers on the dorsal surface, but these feathers were so long that they would cover up the scales in life. One undescribed taxon appears to have some EBFFs, long bristle-like feathers. However, it's not unlikely that some taxa currently grouped in Compsognathidae are not actually compsognathids. There's a paper in the works (or so I've read) that suggests there's only one other known taxon (besides Compsognathus itself) that belongs in Compsognathidae proper, while other "compsognathids" are scattered elsewhere around Coelurosauria or even Tetanurae. So watch this space.
Tyrannosauroid grade (represented by Dilong paradoxus, with inferences from Yutyrannus huali) - Phylogenetic bracketing isn't a lot of help here. Full-body preservation of tyrannosauroid integument is unknown, and a lot of the little bits and pieces we have are still unpublished. Dilong preserves protofeathers or plumulaceous feathers on the dorsal surface of the tail and behind the jaw, while Yutyrannus preserves them on the tail, neck, forelimbs, and feet. Some of the feathers preserved in Yutyrannus and an undescribed tyrannosauroid may be EBFFs. Tyrannosaurid skin impressions from at least the underside of the tail and feet show scales, and patches of naked skin (with very fine scales) are reportedly known (but unpublished) in at least Gorgosaurus libratus and possibly others. It might be worthy to note that some have suggested Dilong paradoxus might not actually be a tyrannosauroid, but most other analyses don't support this so I'll let it slide for now. As you can see, basal coelurosaur phylogeny is still very, very messy.
Therizinosaur grade (represented by Beipiaosaurus inexpectus) - EBFFs and protofeathers or plumulaceous feathers on the neck, legs, and tail. Long protofeathers or plumulaceous feathers on the arms. Face probably mostly naked at least in Bepiaosaurus. Integument elsewhere unknown. Here restored as being fully fuzzy except for the toes based on more derived maniraptors.
Ornithomimosaurs appear to have been similar in distribution and structure of feathers. Feather markings on the forelimb bones of Dromiceimimus brevitertius suggest pennibrachiae (wings) may have been present in the adults, possibly (but not certainly) formed by pennaceous feathers.
Oviraptorosaur grade (represented by Similicaudipteryx yixianensis) - Symmetrical pennaceous feathers all over the body, particularly on the wings and tip of the tail, scales or naked pads possibly present on the fingers, naked on the tip of the upper jaw and beaked on the lower (ceanagnathoids have beaks on both jaws). Usually had primaries and secondaries as adults, except for Caudipteryx zoui, which only had primaries. Juveniles lacked secondaries.
Deinonychosaur and assorted-basal-paravians grade (represented by Microraptor zhaoianus) - Pennaceous feathers all over the body, particularly on the wings, tail, and feet. The tip of the snout was naked. Juveniles had fluffy down instead of fully developed wing feathers. Feathering on the legs variable, often covering at least the metetarsals, with some taxa going down further, sometimes covering even the toes. Taxa without feathers on toes had scales there. Wing feathers asymmetrical in flying taxa, symmetrical in flightless (or near-flightless) ones. Fully flightless taxa had degraded feathers that didn't form a closed vane, similar to the feathers of many modern flightless birds such as struthioniforms. At least one taxon (Microraptor zhaoianus) had feathers on the thumb, akin to an alula. Microraptor also preserves a pair of streamer-like tail feathers in the middle of its tail fan. A feathered crest may have been present in Anchiornis huxleyi, but the one thought to be preserved in Microraptor has been shown to be an artefact caused by crushing. Long pennaceous feathers on tail restricted to the tip in some taxa, present along the entire length in others.
Confuciusornithid grade (represented by Confuciusornis sanctus) - Had primaries and secondaries, naked fingers, and pennaceous feathers all over body. Retained long leg feathers, though not as exaggerated as in more basal paravians. It is unknown if tertials were present. Confuciusornithids had beaked snouts, but these evolved independently from those of modern birds. Some specimens (probably males) had a pair of long ribbon-shaped tail feathers.
Enantiornithine grade (represented by Shanweiniao cooperorum) - Had pennaceous feathers all over body, usually including the snout (which is beaked, half beaked or naked in some taxa). Also had alulas, which more basal maniraptors usually lacked. At least some retained long leg feathers, though not as exaggerated as in more basal paravians. Most didn't have a tail fan, but either had no long retrices at all or a pair of ribbon-shaped tail feathers similar to that of confuciusornithids, though generally not as long. Shanweiniao cooperorum is known to have had a "tail fan" (different from the tail fans of ornithurines), with four ribbon tail feathers instead of two.
Euornithines (represented by Passer domesticus) - Pennaceous feathers all over body, except the beaked snout in neornithines and usually scaly lower legs. The feathering of the legs is variable, however, and ranges from mostly naked to fully feathered. Loss of long leg feathers. Have primaries, secondaries, tertials, alulas, and tail fans.
I have not included the scansoriopterygids, although they do preserve some integument. Only two scansor taxa are known so far. One, Scansoriopteryx heilmanni, had scales underneath its long tail (but I've seen it suggested that these scales are actually from the legs) with downy feathers on the wings, body, and tail. (It's a juvenile, so its adult feathers are unknown.) The other, Epidexipteryx hui, had long ribbon-like tail feathers on its short tail and some pennaceous feathers on the body, but doesn't appear to preserve any wing feathers. Some say that Epidexipteryx might actually be the adult form of Scansoriopteryx, with the tail shortening as the animal reaches maturity. (This isn't as far fetched as it sounds, as the juveniles of some Mesozoic birds, such as confuciusornithids, had longer tails than the adults.) With all these uncertainties, it's difficult to figure out how the integument of scansors were arranged and their exact placement in the coelurosaur tree.