Interestingly, I have run some tests and most of the RF seems to be spraying
off an area
of the aluminum front panel behind which is the frequency setting variable
capacitor.

The large aluminum dials on many RF generators had phosphor-bronze "fingers"
behind them to keep them well grounded. Also, the paint around the front lip of
the case where the panel is screwed on should be scraped off if it isn't
already. The attenuator assembly should be shielded from where the RF enters it
to the output connector.

RF leakage is one of the big differences between service type generators and
professional equipment. The other is, of course, accuracy of calibration.