Not yet reviewed by Pith; the record is open.

The $\sim 3\sigma$ discrepancy between the predicted and observed reactor anti-neutrino flux, known as the reactor anti-neutrino anomaly, continues to intrigue. The recent discovery of an unexpected bump in the reactor anti-neutrino spectrum, as well as indications that the flux deficit is different for different fission isotopes seems to disfavour the explanation of the anomaly in terms of sterile neutrino oscillations. We critically review this conclusion in view of all available data on electron (anti)neutrino disappearance. We find that the sterile neutrino hypothesis cannot be rejected based on global data and is only mildly disfavored compared to an individual rescaling of neutrino fluxes from different fission isotopes. The main reason for this is the presence of spectral features in recent data from the NEOS and DANSS experiments. If state-of-the-art predictions for reactor fluxes are taken at face value, sterile neutrino oscillations allow a consistent description of global data with a significance close to $3\sigma$ relative to the no-oscillation case. Even if reactor fluxes and spectra are left free in the fit, a $2\sigma$ hint in favour of sterile neutrinos remains, with allowed parameter regions consistent with an explanation of the anomaly in terms of oscillations.