sc.pl.paga raises TypeError when cax is passed with multiple colors

Bug report

sc.pl.paga documents cax as "A matplotlib axes object for a potential colorbar" and types it as cax: Axes | None. But when plotting multiple colors, the code indexes it like a sequence (cax[icolor]), which raises TypeError: 'Axes' object is not subscriptable.

Reproduce

import numpy as np, pandas as pd, scipy.sparse as sp
import anndata as ad, scanpy as sc

rng = np.random.default_rng(0)
adata = sc.AnnData(rng.random((60, 20)))
adata.obs['group'] = pd.Categorical(rng.choice(list('abcde'), 60))
k = 5
rows = np.array([0,1,1,2,2,3,3,4,4,0]); cols = np.array([1,0,2,1,3,2,4,3,0,4])
conn = sp.csr_matrix((np.ones(len(rows)),(rows,cols)), shape=(k,k))
adata.uns['paga'] = {'groups':'group','connectivities':conn,'connectivities_tree':conn.copy()}
pos = rng.random((k,2))

fig, ax = plt.subplots()
sc.pl.paga(adata, color=adata.var_names[:2].tolist(), cax=ax, pos=pos, show=False)

Actual

TypeError: 'Axes' object is not subscriptable

Expected

Either accept a single cax (placing the colorbar in it) or a sequence of axes, and document/type accordingly.

Location

src/scanpy/plotting/legacy/_tools/paga.py — cax[icolor] in the colorbar branch.

Run the supplied reproduction first, then inspect src/scanpy/plotting/legacy/_tools/paga.py around the colorbar branch and its cax handling. Add regression coverage for multiple colors with a single Axes, and ensure the documented and typed behavior supports either one axis or a sequence without raising the reported TypeError.