Abstract The hard X-ray source 2PBC J0658.0–1746 was proposed as an eclipsing magnetic cataclysmic variable of the polar type, based on optical follow-ups. We present the first spectral and timing analysis at X-ray energies with XMM–Newton, complemented with archival X-ray, optical, infrared (IR) photometry, and spectroscopy. The X-ray emission shows bright and faint phases and total eclipses recurring every 2.38 h, consistent with optical properties. This firmly identifies 2PBC J0658.0–1746 as an eclipsing polar, the second hard X-ray selected in the orbital period gap. The X-ray orbital modulation changes from cycle-to-cycle and the X-ray flux is strongly variable over the years, implying a non-stationary mass accretion rate both on short and long time-scales. The X-ray eclipses allow to refine the orbital ephemeris with period 0.09913398(4) d, and to constrain the binary inclination $79^{∘}\lesssim i \lesssim 90^{∘}$ and the mass ratio 0.18$\lt M_2/M_{\mathrm{ WD}}\lt $0.40. A companion mass M$_{2}=0.2-0.25\rm \, M_{⊙ }$ with a radius R$_{2}=0.24-0.26\rm \, R_{⊙ }$ and spectral type ∼M4, at D$=209^{+3}_{-2}\rm \, pc$, is derived. A lower limit to the white dwarf mass of $∼ 0.6\, \rm \, M_{⊙ }$ is obtained from the X-ray spectrum. An upper limit to the magnetic colatitude, $β \lesssim 50^{∘}$, and a shift in azimuth, $ψ ∼ 14^{∘}$, of the main accreting pole are also estimated. The optical/IR spectral energy distribution shows large excess in the mid-IR due to lower harmonics of cyclotron emission. A high-state mass accretion rate $\rm \, ∼ 0.4-1\times 10^{-10}\, M_{⊙ }\, yr^{-1}$, lower than that of cataclysmic variables above the gap and close to that of systems below it, is estimated. With 2PBC J0658.0–1746, the number of hard X-ray-selected polars increases to 13 members, suggesting that they are not as rare as previously believed.