Note that these commands are not efficient to be run routinely on gigantic datasets, instead are only for quickly eyeballing and inspecting relatively smaller datasets.
# print the raw-signal (column 8) for the read-id r1:
slow5curl get --to slow5 https://url/to/reads.blow5 "r1" | grep -v '^[#@]' | awk '{print $8}'
# extract the signal samples 100-200 (1-indexed) for the read-id r1:
slow5curl get --to slow5 https://url/to/reads.blow5 "r1" | grep -v '^[#@]' | awk '{print $8}' | cut -d, -f 100-200