Datetime formats inconsistent with the data

In the schema the datetime formats are "fmt:%Y-%m-%dT%H%M%SZ". For example:

{
    "name": "utc_timestamp",
    "description": "Start of timeperiod in Coordinated Universal Time",
    "type": "datetime",
    "format": "fmt:%Y-%m-%dT%H%M%SZ"
},

but data in time_series_15min_singleindex.csv has

2014-12-31T23:00:00Z,2015-01-01T00:00:00+0100,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
2014-12-31T23:15:00Z,2015-01-01T00:15:00+0100,,,,,,,,,,37248.0000,,,27913.0000,,,667.0000,,,27246.0000,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

Note the colons between hours, minutes and seconds.

I also wonder why the ’fmt:’ prefix. At least pandas to_datetime() function seems to be unable to use the format string with this prefix.