The problem with the 5 HZ is that of settling time. i.e. how long it will take for the output to settle at its correct level and frequency. You do not say how fast the change of frequency will be, i.e. 5-150 HZ in one second, 10 seconds. . . or how accurate the sine waves must be i.e. less then .01% THD, 1%?, 3 %?
One way to make a precision generator would be to have a fixed 100 KHZ oscillator , mix this with a variable 100,005 -> 100,150 HZ oscillator, filter the results using a LPF with a cut off at 50K HZ will give you the basic 5- 150 HZ variable. using a sniff of the oscillators, multiply them both by X6 ( X3 first then X2), a second mixer, identical LPF will give the X6 output. Because the change of frequencies are so low compared to the RF frequencies they are generated at, the levels will be flat. The mixers should give you less then .1 % THD.
Frank