The progenitor of SN 2005cs, in the galaxy M51, is identified in pre-explosion Hubble Space Telescope (HST) Advanced Camera for Surveys (ACS) Wide Field Channel (WFC) imaging. Differential astrometry, with post-explosion ACS High Resolution Channel (HRC) F555W images, permitted the identification of the progenitor with an accuracy of 0.006 arcsec. The progenitor was detected in the F814W pre-explosion image with I= 23.3±0.2, but was below the detection thresholds of the F435W and F555W images, with B < 24.8 and V < 25 at 5σ. Limits were also placed on the U- and R-band fluxes of the progenitor from pre-explosion HST Wide Field Planetary Camera 2 (WFPC2) F336W and F675W images. Deep images in the infrared from the Near Infra-red Imager (NIRI) on the Gemini North Telescope were taken 2 months prior to explosion, but the progenitor is not clearly detected on these. The upper limits for the JHK magnitudes of the progenitor were J < 21.9, H < 21.1 and K < 20.7. Despite having a detection in only one band, a restrictive spectral energy distribution of the progenitor star can be constructed and a robust case is made that the progenitor was a red supergiant with spectral type between mid-K and late-M. The spectral energy distribution allows a region in the theoretical Hertzsprung-Russell diagram to be determined that must contain the progenitor star. The initial mass of the star is constrained to be MZAMS= 9+3–2M☉, which is very similar to the identified progenitor of the Type II-P SN 2003gd, and also consistent with upper mass limits placed on five other similar supernovae. The upper limit in the deep K-band image is significant in that it allows us to rule out the possibility that the progenitor was a significantly higher mass object enshrouded in a dust cocoon before core collapse. This is further evidence that the trend for Type II-P supernovae to arise in low- to moderate-mass red supergiants is real.