underclipping after an insertion

[bwlang]

At one locus in this dataset I see that iVar is soft-clipping all the bases except 1 when that base is next to an insertion and mismatches the reference.
I've also included illumina data from the same sample so the truth is clear. Compounding the problem, there does seem to be a systematic insertion problem with the ONT reads at this position (visible in the second screenshot from both amplicons). There is no evidence of an insertion in the illumina sequences.

I don't think this is the same issue as #86 or #85 since it's not really about deletions, nor is it related to alignment issues near ends of reads.

To Reproduce
I reproduced the full analysis in this galaxy history so you can see all the details of how it was run.
https://usegalaxy.org/u/bradlanghorst/h/sars-cov-2-saliva

Expected behavior
iVar should clip the base just after the insertion that now overlaps with the primer sequence

Screenshots

[bwlang]

thanks to @wm75 and @mvdbeek for the helpful comments

[bwlang]

looks like the link URL is no longer valid.
Here's an updated link:
https://usegalaxy.org/u/bradlanghorst/h/nebnext-artic-true-positive-saliva-illumina--ont