We report on the detection of a giant radio halo in the cluster Abell 3404 as well as confirmation of the radio halo observed in Abell 141 (with linear extents $\sim\!770$ and $\sim\!850$ kpc, respectively). We use the Murchison Widefield Array, the Australian Square Kilometre Array Pathfinder, and the Australia Telescope Compact Array to characterise the emission and intervening radio sources from $\sim100$ to 1 000 MHz; power law models are fit to the spectral energy distributions with spectral indices $\alpha_{88}^{1\,110} = -1.66 \pm 0.07$ and $\alpha_{88}^{943} = -1.06 \pm 0.09$ for the radio halos in Abell 3404 and Abell 141, respectively. We find strong correlation between radio and X-ray surface brightness for Abell 3404 but little correlation for Abell 141. We note that each cluster has an atypical morphology for a radio-halo-hosting cluster, with Abell 141 having been previously reported to be in a pre-merging state, and Abell 3404 is largely relaxed with only minor evidence for a disturbed morphology. We find that the radio halo powers are consistent with the current radio halo sample and $P_\nu$–M scaling relations, but note that the radio halo in Abell 3404 is an ultra-steep–spectrum radio halo (USSRH) and, as with other USSRHs lies slightly below the best-fit $P_{1.4}$–M relation. We find that an updated scaling relation is consistent with previous results and shifting the frequency to 150 MHz does not significantly alter the best-fit relations with a sample of 86 radio halos. We suggest that the USSRH halo in Abell 3404 represents the faint class of radio halos that will be found in clusters undergoing weak mergers.