Your panel is intended to provide a trickle charge current into a lead acid battery (nominal float voltage 13.8V) as cheaply as possible. That's nearly three times more voltage than you need for USB devices. To get as most power out of it as possible without the complexities of a MPPT controller, you need to run the panel into a constant voltage load at approximately 75%-80% of its nominal O/C terminal voltage. That's a fairly good match to a SLA battery, and it sounds like there is little risk of overcharging, so you'd only need an under-voltage cutout, set fairly high so the SLA battery itself isn't killed by repeated cycling to below 50% charge.
If you want to eliminate the SLA battery, that would involve hacking the phone charger DC-DC converter's feedback loop to drop the nominal 5V output voltage proportionately as the panel output drops from around 15V to 13V, and *IF* there is enough bulk capacitance on the input to the DC-DC converter to smooth out any pulsing due to the load's charging algorithm, it will then settle at a current and voltage that can be sustained. This may require a very large reservoir cap if the load has a 'smart' charge controller.
If you lift the top end of the lower resistor in the MC34063A feedback potential divider and add a small N channel MOSFET with a Vgs rating >25V (or use a 2n7000 with an 15V Zener from gate to ground to protect it) between the feedback pin (D) and the lower resistor (S), with a potential divider feeding its gate from the input voltage (I suggest a 220K upper resistor and a 100K pot for the lower resistor), it should do the job. Start with the pot at max so the extra circuit has virtually no effect (MOSFET hard on), load the 5V rail with resistors till the voltage drops to about 10V, then tweak the pot to get it back up to about 13V. You can do the same thing with a bipolar transistor, but it needs to go between the lower feedback resistor and ground, and obviously it needs different resistors in the divider.