We recently presented evidence that stable umbra-penumbra boundaries are characterised by a distinct canonical value of the vertical component of the magnetic field, $B^{\rm stable}_{\rm ver}$. In order to trigger the formation of a penumbra, large inclinations in the magnetic field are necessary. In sunspots, the penumbra develops and establishes by colonising both umbral areas and granulation, that is, penumbral magneto-convection takes over in umbral regions with $B_{\rm ver} < B^{\rm stable}_{\rm ver}$, as well as in granular convective areas. Eventually, a stable umbra-penumbra boundary settles at $B^{\rm stable}_{\rm ver}$. Here, we aim to study the development of a penumbra initiated at the boundary of a pore, where the penumbra colonises the entire pore ultimately. We have used Hinode/SOT G-band images to study the evolution of the penumbra. Hinode/SOT spectropolarimetric data were used to infer the magnetic field properties in the studied region. The penumbra forms at the boundary of a pore located close to the polarity inversion line of NOAA\,10960. As the penumbral bright grains protrude into the pore, the magnetic flux in the forming penumbra increases at the expense of the pore magnetic flux. Consequently, the pore disappears completely giving rise to an orphan penumbra. At all times, the vertical component of the magnetic field in the pore is smaller than $B^{\rm stable}_{\rm ver} \approx 1.8$~kG. Our findings are in an agreement with the need of $B^{\rm stable}_{\rm ver}$ for establishing a stable umbra-penumbra boundary: while $B_{\rm ver}$ in the pore is smaller than $B^{\rm stable}_{\rm ver}$, the protrusion of penumbral grains into the pore area is not blocked, a stable pore-penumbra boundary does not establish, and the pore is fully overtaken by the penumbral magneto-convective mode. This scenario could also be one of the mechanisms giving rise to orphan penumbrae.