Where the headland narrows, a great earthen bank and fosse, a defensive ditch, were cut roughly 75 metres across the neck of the promontory, sealing off the seaward section from the fields behind. The inlet of Coosgorm bites deep into the southern flank, further thinning the connection to dry land. The result is a fort that uses the sea as most of its wall, with only that one engineered barrier doing the work of the rest.

The antiquarian T. J. Westropp visited in 1910 and recorded the site under the name Dunsheane, though the promontory is more broadly known as Doonmore, with the smaller and now largely inaccessible projection on the far side of Coosgorm called Doonbeg. The earthen bank survives in considerable bulk: 10 metres wide at the base, rising 3 metres on the interior side and a more formidable 6.6 metres above the base of the fosse. Along its top, large horizontally-laid slabs form an intermittent stone wall about 1.5 metres wide. Access at the western end is through a causewayed entrance 4 metres wide, while Westropp noted what he called an old-looking gateway at the eastern end, where the stumps of two pillar stones still stand 1.5 metres apart. He also recorded a souterrain near the inner base of the bank, roughly 10 metres east of the causeway. A souterrain is an underground stone-lined passage, typically associated with early medieval settlement and used for storage or refuge. That passage is no longer accessible, its presence now signalled only by a hollow in the ground and some settled subsidence in the bank above. Within the interior, circular depressions cut back into the sloping ground mark the positions of up to five possible hut-sites, ranging from 3 to 5.5 metres across. Three more that Westropp noted along the north-eastern cliff edge could not be located during later survey work.

Doonbeg, the separate triangular spur beyond Coosgorm, is connected to Doonmore by a narrow isthmus that drops away almost sheer in places. When Westropp visited, a fosse was still visible across the grassy slope at what he described as the saw-like remnant of the neck. Whether that fragment is still legible on the ground today is unclear, though the terrain alone makes any approach to it a serious undertaking.