Gibson Assembly is not ideal for short fragments; chances are that the T5 Exonuclease will digest your entire fragment before it has the chance to hybridize with the backbone. For fragments shorter than 200 bp NEB recommends a 5-fold excess to compensate for this, but in your case the fragment would only be around 130 bp long. It might work, but efficiency will be very low.
Either way, based on my understanding that you don't have the 3xtag-TEV sequence available as PCR template, what you would need is two long primers designed such that they overlap by 20-25 nt at their 3' ends (orange), and with the backbone at their 5' ends (blue) as shown below:
You would then run PCR cycles with the primers alone to generate your double-stranded fragment, which you can then use in the Gibson Assembly reaction.
If you're not using a commercial Gibson Assembly kit, but instead add the enzymes yourself, you can also try an alternative method called TEDA cloning. The procedure is practically the same as that of Gibson Assembly, but TEDA uses only T5 Exonuclease. Instead of relying on (very expensive!) Phusion DNA Polymerase and Taq DNA Ligase for sealing the plasmid, plasmid repair in TEDA happens after transformation into bacteria, taking advantage of E. coli's own DNA repair machinery.
In your case, the advantage with TEDA is that you can try different incubation times with T5 Exonuclease to find a short enough incubation time such that single-stranded 5' homology ends are generated without digesting your entire fragment; or even incubate your backbone first and add your short fragment later in the TEDA reaction. As an extra bonus, TEDA cloning costs a fraction of the price of a Gibson Assembly reaction.