Ignition or other electrical systems are responsible for 90% of engine performance problems. It's hard to say how it could still be haywire, with two good mags, new harness and sparkplugs. That's why I had suggested the P-leads.
Reducing the throttle tends to lean the mixture a bit all by itself. The mixture control takes it further. A stumble during either of those implies a weak spark.
What are those mags? Bendix or Slick? Some Bendixes had an extra condenser on the outside of the mag to suppress ignition noise, and it's wired to the P-lead terminal, which puts it in parallel with the condenser inside the mag. That has the effect of increasing the overall capacitance, which can weaken the spark. It absorbs too much of the primary flow after the points open before the primary flow finally stops and the flux collapses. It makes the collapse later and weaker. If it's there you might try going without it. Or maybe someone has put noise filters in the P-lead lines on the firewall.
There are ADs on older Bendix mag switches and all ACS/Gerdes mag switches. The Bendix one is a safety issue--the switch might not ground both mags when it's shut off--and the ACS AD involves burned contacts inside the switch that can cause mag issues. The start contacts of the switch are right up against the mag contacts, and debris can make trouble. You have to crawl under the panel to see which you have. Don't trust the AD signoff your mechanic might have made; I have found many of them signed off as "Bendix mag switch tested and found OK" when the switch was actually an ACS. Many mechanics don't seem to know that there are more than Bendix switches out there. Cessnas are full of ACS switches.
https://rgl.faa.gov/Regulatory_and_...B8ABD56539B4684886256A3E00759DBF?OpenDocument