Temnospondyls apparently had smooth, unarmored, skin.
I don't know where the surprise is.
There are two entirely different entities that are called "scales".
Meanwhile, it's reassuring to hear that some stem-group amphibians seem to have shared a derived soft-tissue condition with the crown group.
That is not the case.
The on-going [sic] controversy over lissamphibian origins is fueled by their highly derived morphology with respect to archaic fossil forms. Here we report a new specimen that bridges the gap between amphibamid temnospondyls and basal frogs and salamanders. The specimen is missing the zeugopods, most of the manus, and the ventral-most portion of the skull. The skull is broad and rounded, but it has a lighter, more strut-like construction than is typical for temnospondyls. The extremely large interpterygoid vacuities are bordered by a pterygoid that just fails to reach the lateral skull margin except by a dorsal projection, a palatine, and a narrow vomer that bears denticles arranged in three rows on a ridge along the medial margin of the large choana. Marginal teeth are tiny, monocuspid, pedicellate cones. A very large otic notch, with articulation scars for the tympanic annular cartilage, closely approaches the orbital margin and creates a narrow postorbital bar. There are 17 presacral vertebrae, and the caudal vertebrae are poorly ossified and rapidly taper out. Ribs are short, laterally projecting elements with spatulate distal tips. The olecranon is ossified. A basale commune is present in the pes, which has a phalangeal formula of ?-2-3-4-3. This specimen has a mosaic of amphibamid characters and synapomorphies of both frogs and salamanders. The overall impression of the skull is frog-like, except the frontals are not coossified with the parietals. Pedicellate teeth are [among tetrapods] known only from lissamphibians, amphibamids, and possibly one branchiosaur. The vomers, with the lack of fang-pit pairs and [the] raised patch of denticles in rows[,] are especially batrachian. The vertebral count is transitional between *Amphibamus* (21) and *Triadobatrachus* (14). The tail appears to be in the process of being lost and the phalangeal count suggests that this specimen may be the most basal frog; however, the basale commune is unique to salamanders. The presence of a basale commune in an amphibamid suggests that preaxial digital development was more widespread among batrachian sister[-]groups, and may have been primitively present in frogs.
The paper should be out soon.
Lissamphibians are one of the last clade[s] of extant tetrapod whose origin in the fossil record is still currently very [much] debated. Three main hypotheses are [being] discussed: a monophyletic origin within dissorophoid temnospondyls, a monophyletic origin within lepospondyls and a polyphyletic origin with urodeles and anurans evolving from two clades of dissorophoid temnospondyls[*] and caecilians from microsaur[ian] lepospondyls. The anatomy of two lepospondyls have [sic] been studied. The late Permian diplocaulids from Morocco are redescribed and the systematics of nectrideans is discussed. Moreover, they are the last lepospondyls known in the fossil record. An aistopod from the Upper Carboniferous of France is also described. It exhibits some anatomical features that were previously known only among Lissamphibians [sic] and some dissorophoids. This study reveals that some synapomorphies used to unit[e] Lissamphibians and temnospondyls are homoplas[t]ic. A developmental study of ossification sequences of urodeles reveals that these developmental sequences are highly variable and should be used very carefully to solve the problem of lissamphibian origin. Finally, a phylogenetic analysis of a large scale of Paleozoic tetrapods is performed and shows that each hypothesis concerning the origin of Lissamphibians is weakly supported. The main causes of these incongruences are various. An extendsive homoplasy reigns among Paleozoic tetrapods. The lack of "transitional" fossils provides huge morphological gaps between crown-lissamphibians and their putative sister-taxa and temporal gaps in the beginning [of] the Carboniferous (the [sic] Romer's gap) where all groups of Paleozoic tetrapods diversified, with a very poor fossil record, and in the Permo-Triassic boundary, after which the first Lissamphibians appeared whereas all their putative sister-groups had disappeared well before. Some new axes of research are thus proposed in order to solve the origin of Lissamphibians despite the lack of fossils that could exhibit enough features to anchor Lissamphibians in one (or more) Paleozoic group.