A framework for unified analysis of small-signal and large-signal power system stability based on static and dynamic eigenvalues is proposed in this paper. The presented implementation is based on Gear's method, which is a two-step integration method for numerical simulation with self-adaptive time-step. Furthermore, it can be easily configured for providing the state matrix as basis for calculating the system eigenvalues during simulation. Thus, the presented framework allows for eigenvalue-based analysis of small-signal dynamics and stability margin at any steady-state operating point during a time-domain simulation. Furthermore, Linear Time-Varying system theory is utilized for modal analysis during large-signal transients. For this purpose, dynamic eigenvalues and eigenvectors are calculated by solving a Riccati equation to generalize the modal analysis during transient conditions. The stability is evaluated by calculating the Lyapunov exponent of the mode-vector of the system. The results from numerical analysis of three case studies are presented to evaluate and illustrate the characteristics of the presented approach for unified small-signal and transient stability analysis. © 2022 The Authors. IET Generation, Transmission & Distribution published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.