Convert the latest TRF output, using the '-ngs' flag, to gff. Very rudimentary, use at your own risk.

trf2gff.py

with open(fname) as fh:
    for line in fh:
        ele = line.strip().split(" ")
        if line.startswith('@'):
            seq_name = ele[0][1:]
        else:
            [start, stop, period, copies, 
             consensus_size, perc_match, perc_indels, 
             align_score, perc_A, perc_C, perc_G, perc_T, 
             entropy, cons_seq, repeat_seq, left_flank, right_flank] = ele
            gff_line = [seq_name, 'TRF', cons_seq + '_' + copies + '_copies',
                        start, stop, '.', '.', '.', 'Name='+ cons_seq + '_' + copies + '_copies']
            print '\t'.join(gff_line)