I like to start from C^4 instead of C^5 because in that way the structure is very much as spacetime, signature (1,3). And 1-3= 6 mod 8
The 0, 1, 2, 3, and 4 forms (The Clifford algebra, if you prefer) are generated from a charged but uncoloured and three coloured generators:
\nu = .
e^+= dt
d_r = dx
d_g = dy
d_b = dz
u_r = dt \wedge dx
u_g = dt \wedge dy
u_b = dt \wedge dz
\bar u_b = dx \wedge dy
\bar u_g = dx \wedge dz
\bar u_r = dy \wedge dz
e^-= dx \wedge dy \wedge dz
\bar d_r = dt \wedge dy \wedge dz
\bar d_g = dt \wedge dx \wedge dz
\bar d_b = dt \wedge dx \wedge dy
\nu = dt \wedge dx \wedge dy \wedge dz
This idea is based on Unified Theories For Quarks And Leptons Based On Clifford Algebras by R. Casalbuoni (CERN) , Raoul Gatto (Geneva U.) . UGVA-DPT 1979/11-227, Nov 1979 Published in Phys.Lett.B90:81,190 and Families from Spinors by Frank Wilczek , A. Zee . Phys.Rev.D25:553,1982.
The cap product by the volume form maps particle to antiparticle, or almost. Chyrality considerations pending, a volume form seems very much as a mass term (or a higgs term)
From two copies (left and right) of it, you build the C^5 Baez is speaking about. And
it is possible to built the C^4 thing from two copies in C^3, using only the coloured generators. In that way it is very close to Harari-Shupe.