I think there is not enough data to provide an exact answer theoretically.
You can figure out an estimate value by virtually titrating the tripeptide and figuring out after what step an isolectric point likely situates:
$$
\def\VEQ{{\scriptsize-\ce{H+}}{\Large\downharpoonleft\!\!\upharpoonright}\scriptsize{+\ce{H+}}}
\begin{array}{rcc}
\mathrm{pH} < 2.2 &\quad &\ce{H3\overset{+}{N}-\overset{+}{Lys}-\overset{+}{Lys}-\overset{+}{Lys}-COOH} \\
& & \VEQ \\
2.2 < \mathrm{pH} < 8.0 &\quad &\ce{H3\overset{+}{N}-\overset{+}{Lys}-\overset{+}{Lys}-\overset{+}{Lys}-COO-} \\
& & \VEQ \\
8.0 < \mathrm{pH} < 10.5 &\quad &\ce{H2N-\overset{+}{Lys}-\overset{+}{Lys}-\overset{+}{Lys}-COO-}\\
& & \VEQ \\
10.5 < \mathrm{pH} = \mathrm{pI} &\quad &\ce{H2N-\overset{+}{[Lys-Lys-Lys]}-COO-}\\
\end{array}
$$
As you can see, the formal net zero charge doesn't occur between any given $\mathrm{p}K_\mathrm{a}$s.
Rather, at the highest $\mathrm{p}K_\mathrm{a}(\ce{Lys{-}ε{-}NH3+}) = 10.5$ the tripeptide is still going to have $ 0.5 + 0.5 + 0.5 -1 = +0.5$ formal charge considering $50\,\%$ deprotonation of residue amine groups at $\mathrm{pH} = \mathrm{p}K_\mathrm{a}$ and $-1$ charge at the $\ce{-COO^-}$ group.
This implies that $\mathrm{pI}$ of trilysine must be slightly higher, about $10.8:$

For polylysines $\mathrm{pI}$ should increase with the chain length.
ThermoFisher's Peptide Analyzing Tool suggests theoretical $\mathrm{pI}$ values of $11.0,$ $11.2$ and $11.3$ for tetra-, penta- and heptalysine, respectively.
This tool, however, refuses to calculate isoelectric points for tripeptides (Input sequence is too short
).
This can be an indication of insufficient quality of applied algorithm resulting in apparently wrong data suggested by Swiss-Prot in your question.
Note that we didn't corrected $\mathrm{p}K_\mathrm{a}$ values for the tripeptide.
I wouldn't assume they change drastically, but it's not the identical values either.