The main problem here is that neither branch of the manuscript tradition supplies an acceptable text for the word or words between quosdam and nactus. But an additional worry is that none of the solutions on offer enables one to give a satisfactory account of the transmission. Suis is particularly hard to account for.
Stephanus' suggestion, consili sui, takes its cue from two references to a plan, presumably one common to Pothinus and his ally Achillas, in the subsequent paragraph (3.109.2 relinquebatur ut … consilium … Achillae cognosceret [sc. Caesar], 3.109.6 priuato paucorum et latronum quam regio consilio). It suits the context and produces acceptable Latin. If the text was originally consili sui, one can assume either a copying error (-ili ⟶ -scii, perhaps under the influence of the conspiracy suggestion in clam or the Ciceronian pairing of adiutor and conscius) or a content-based gloss 'conscii' that ousted consili from the text. The archetype will then have read conscii sui, to which mu applied a rough-and-ready repair resulting in conscios sui, an expression that was tolerable in a later era, at least, albeit with a meaning more like that of classical conscius sibi (TLL 4.372.53–5). Nu transmitted conscii, and S, as so often, aligned its ending with that of the following word. So far so good. But what is responsible for suis? (V shows its discomfort with this reading by leaving su without an ending and writing is as a separate word.) It cannot be a deliberate innovation, and there is no obvious spur for it as an unconscious error. Nor does it seem to be related to the other distinctive readings of nu here. (The omission of ex and the tense of praeficit look archetypal' pg 307since the variants in mu are easy to understand as innovations designed to fix minor grammatical problems.) The story is marginally better if one assumes that the archetype read conscii suis, which nu transmitted without alteration and mu adapted, cleaning up the sentence here as in the other two spots. But this means that in the archetype we see unrelated innovations in adjacent words.
Apitz's suggestion, conscios <consili> sui, also produces good sense and good Latin (cf. V. Max. 7.4.5 'si huius consili mei interiorem tunicam consciam esse sensero, continuo eam cremari iubebo'). But if it is right, the transmitted readings make even less sense. One assumes that consili has fallen out of the text through haplography after conscios, and that mu transmits the archetype's corrupt text. Fine. But then nu makes two nonsensical and incompatible innovations.
Another approach is to delete both words as a gloss that made its way into the archetype's text and was transmitted by mu with a minimal repair and by nu with a nonsensical innovation. This explanation is no more satisfying than the others.
The text I print, without much conviction, assumes that the archetype read conscii suis and that both words belong to an original conscios, of which the beginning and ending somehow got separated, whereupon the ending was transmuted into an independent word. (For the absolute use of conscius, which is widespread in Cicero and others, see TLL 4.370.32–69; the best parallel for our passage is Clu. 36 non modo conscio sed etiam conuiua et adiutore Oppianico.) Conscios contributes more than the neutral consili sui to the negative characterization of Pothinus, whose execution is the last event of Caesar's narrative (3.112.12), for what that is worth. (Not much.) The elimination of sui(s) also sharpens the antithesis between inter suos in the preceding sentence (is primum inter suos queri atque indignari coepit regem ad causam dicendam euocari) and the focus on the regis amici in the present one, an antithesis structured by primum… deinde. Again, far short of decisive. But the idea that Pothinus operates independently of the king's wishes and perhaps interests recurs with an emphasis that troubles the text at 3.108.3 (hunc [sc. Achillam] incitatum suis et regis inflatum pollicitationibus quae fieri uellet litteris nuntiisque edocuit), and again emphatically at 3.109.6 (priuato paucorum et latronum quam regio consilio). And his ally Achillas initiates a murderous attack on the king's emissaries and necessarii, Dioscorides and Serapion (3.109.4–5).