Three concentric limestone walls enclose a total area of 0.81 hectares, but along the north-western side the outer and middle walls simply stop at the cliff face, letting the drop itself serve as the final barrier. It is that detail, the deliberate incorporation of a sheer ravine into a defensive scheme, that gives the place its particular character. The innermost wall, at its thickest point eight metres wide on the eastern side, was built with a rubble core faced on both surfaces with coursed limestone flags, and its entrance was a paved pathway flanked by sidewalls. Five niches were once visible on the interior face of that inner wall; only two remain today.

The site was excavated in 1934 by Hugh O'Neill Hencken as part of the Third Harvard Archaeological Expedition to Ireland, and what he found was a settlement of some complexity. He identified buildings for a chief occupant, those second in command, a guardhouse, kitchens, and servants' quarters, and calculated that the fort could have housed forty to fifty people. Two souterrains, underground stone-lined passages typically used for storage or refuge, were found beneath the interior buildings, each reached by a flight of steps. One of them yielded a silver annular brooch from a layer of ash on its floor; the condition of the piece suggests it was deposited late in the ninth century or early in the tenth. Datable metalwork across the site spans as much as five centuries, from the fifth to the ninth, though the question of when the cashel was actually built, and whether earlier activity preceded it, has not been settled. Stone axes, saddle querns, and flint scrapers were also recovered, but without stratigraphic links to the walls, they may belong to a phase that predates the monument entirely. Perhaps the most unsettling discovery was scattered human bones throughout the habitation layers, mixed into kitchen debris, many of them broken, and some showing signs of having been chopped. The outermost wall preserves nine vertical joints along its circuit, which Hencken interpreted as the seams between the work of different construction teams, a rare trace of the labour itself still legible in the stone.