A decent wattmeter, like a Bird with a PEP adapter, a 'scope with response to 50 MHz, a counter with an external 'probe' input and a decent signal generator should get you by.
A stable, precise 1 kHz tone generator with a speaker on it is the easiest way to set SSB transmitters on frequency. Had to build ours, not sure where to send you for that item.
The signal generator is the sticky item. It needs to be stable, and have a calibrated output. Most "budget" RF signal generators are not stable enough to use with a SSB receiver, and the calibrated output attenuator is an expensive item all by itself. An under-$200 type will almost never include that, but an UN calibrated output. Many of those will 'leak' several microVolts of output turned down to "zero".
Bought a Boonton 102C sig. gen. on Ebay a few months back, didn't think to ask why the display was dark in the picture. Found out when it arrived. Internal shield covers were missing, the fuse was blown, and the tops blown (literally) off the tops of chips in the power-supply regulator circuits. Now I know never to bid on anything that isn't fully "lit up" in the auction photo.
The venerable, old Heathkit "Cantenna" dummy load has a useful feature for AM/SSB radios. A signal tap-off that's attenuated and rectified with a signal diode. It's meant to be used with a DC voltmeter, to read relative output power. By changing the .01 uf disc cap across the DC output jack to a .001 or so, and shunting 10k across it, a patch cord to the 'scope input will display the modulation envelope. Nearly any cheap scope, no matter the rated vertical frequency response can be used to observe the output of a transmitter into the dummy, since all that comes out of that jack is the demodulated audio, riding on a DC voltage in proportion to the carrier level.
A high-precision, stable synthesized signal generator is a big plus, but not cheap.
The single most-used tool here is a so-called "component tester". I'm told that in the Navy technical command, it was nicknamed the "octopus". It's a two-terminal curve tracer that feeds (on our unit) 2 Volts peak-to-peak AC out the two wires to the probe. The vertial axis on the scope is current, the horizontal is voltage. With no circuit across the probes, you get a flat baseline, all voltage no current. With the probes shorted, you get a vertical line (all current, zero voltage). A semiconductor P-N junction will show a "dog leg" trace, flat on the reverse-bias side of zero, and a vertical "leg" about 6/10 of a volt to the right of center. Mighty handy for identifying blown or damaged junctions in transistors or diodes. Since it's much faster than a meter, you can see the "noise" on a scratchy or intermittent switch when it is opened and closed. A pot with a bad spot will be immediately evident when it's turned.
Several kinds of oscilloscope from MCM, Beckman Industrial, and Elenco have been sold with this "component tester" feature built-in. Most of those place 10 or 20 Volts P-P AC across the probe tips. It places a forward-biased junction "knee" much closer to the baseline's center point. They will permit you to see the "knee" of a low-voltage zener diode, where our 2-Volt version will not.
It's an overlooked kind of tool, but has been an enormous time-saver for us. A commercial version called the Huntron Tracker has been on the market for a while, but the prices are out of reach for me.
Best of luck on getting set up. One more little item that I have found very useful is a 'scope that has a sample of the vertical channel available on a "channel 1 output" socket. Higher-ticket Tek analog 'scopes come with that feature. We patch it into a frequency counter. Turns the 'scope's vertical channel into a broadband preamp for the frequency counter. Makes it possible to use a times-ten 'scope probe to read internal frequencies in the radio that would be too weak for the counter to read with the probe plugged directly into it. Touching a probe to a crystal oscillator circuit will "load" the circuit since there is always capacitance to ground inside the probe. When you remove the probe, the frequency will shift from what it was while you were reading it. Using a 'scope's input channel as a preamp allows touching the probe to the metal case of a crystal. So long as there is no ground wire on the crystal's case, you will pick up enough of the crystal's waveform with the vertical sensitivity turned up, and get a reading with nearly no disturbance to the setting of trimmer caps.
Well, that's a little more than two cents' worth, maybe a quarter.
Best of luck and 73