We seek Clifford numbers c such that φ(c) = cvc–1 leaves all elements v of V fixed, i.e. v = cvc–1 (↔ cv = vc). The only such numbers I know now are the non-zero reals within Cl(n). We restrict ourselves to values v such that (ip v v) = 1. It suffices to require that b = cbα(c)–1 for each base element b of V.

An alternate definition of the Clifford group uses the expression cvα(c)–1 in place of cvc–1. This defines the same set of Clifford numbers but the orthogonal matrices they lead to have the opposite sign when c is odd.

To find those Clifford numbers c that commute with each member of V we first seek all c such that cγ0 = γ0c.
Split c thus c = f + γ0g where neither f nor g mention γ0.
0 = γ0c ↔ (f + γ0g)γ0 = γ0(f + γ0g) ↔ fγ0 + γ00 = γ0f + γ0γ0g ↔
γ0f = fγ0 & γ00 = γ0γ0g ↔ f and g are both pure even. (Note that γ0 has an inverse: –γ0.)
The code corroborates this.

Said another way, a Clifford number commutes with γ0 iff each of its terms mentions an even number of other base vectors. When we require that a Clifford number commute with each of the base elements we are clearly left with only the reals among the Clifford numbers.