Super-resolution microscopy normally requiring complex and cumbersome optics is not applicable for in situ imaging through a narrow channel. Here, we demonstrate single hair-thin multimode fiber (MMF) endoscope (less than 250 $μm$) for in vivo light-field nano-imaging, which is called spatial-frequency tracking adaptive beacon light-field encoded nano-endoscopy (STABLE nano-endoscopy) that enables three-dimensional (3D) subcellular-scale imaging. Spatial-frequency tracking provides up to $10^3$ Hz disorder tracking that ensures stable imaging in long-haul MMFs (up to 200 m) under various conditions. Full-vector modulation and fluorescence emission difference are combined to enhance the imaging signal-to-noise ratio two times and to improve the resolution to sub-diffraction-limited 250 nm ($λ/3NA$). STABLE nano-endoscopy and white-light endoscopy (WLE) are integrated to achieve cross-scale in vivo imaging inside the lumen. This high-resolution and robust observation in a minimally invasive manner paves the way to gain a deeper understanding of the disease mechanisms and to bridge clinical and biological sciences.