fix: check query_time instead of response_time_s list in xlsx output

[imxde-code]

Summary

Fixes #2891

The condition on line 906 checks if response_time_s is None, but response_time_s is initialised as an empty list ([]) on line 896 of the same block — so this check is always False and the empty-string fallback is never reached.

As a result, when a site's query_time is None (timeout or error), the raw None gets appended to the list. Pandas then writes None into the Excel cell instead of a blank string, which breaks downstream xlsx processing.

Root cause

# Before (line 906) — condition always False, None leaks into the list
if response_time_s is None:
    response_time_s.append("")
else:
    response_time_s.append(results[site]["status"].query_time)  # appends None

Fix

# After — checks the actual value, mirrors the CSV path (lines 875-877)
if results[site]["status"].query_time is None:
    response_time_s.append("")
else:
    response_time_s.append(results[site]["status"].query_time)

This mirrors the pattern already used correctly in the CSV output block (lines ~875–877).

Test plan

• Run sherlock --xlsx <username> — sites that time out now produce blank cells instead of None
• Existing tests pass