As power systems transition toward configurations with increasing shares of inverter-based resources (IBRs), transient stability assessment becomes more challenging. Conventional time-domain simulations provide high-fidelity results but are computationally expensive and typically yield binary stable/unstable classifications. To support applications requiring fast transient stability assessment, screening methods based on transient energy functions, extended equal-area criteria, and related direct methods have been proposed. Their main advantage is computational speed, achieved through significant model simplifications. While these methods are well suited to simplified contingency screening, the required simplifications can limit their applicability to planning and detailed system analysis. This work adopts and extends classical metrics developed for synchronous generators, specifically center-of-inertia-corrected dot products and rotor-angle coherency, for transient stability assessment of IBR-rich power systems. These metrics are evaluated using the full implemented model without network reduction or the model simplifications associated with direct methods. They capture the relationship between power imbalances, speed deviations, and angular dispersion and are combined into a Composite Severity Index. Based on short post-fault responses, these metrics transform multivariable trajectories into quantitative measures of transient severity, enabling contingency ranking and providing a continuous stability metric that can be incorporated into optimization frameworks to quantify system-level transient stability requirements.