Collapse Index (CI) is a lightweight, domain-agnostic framework designed to quantify instability in artificial intelligence systems through perturbation-based analysis.This report demonstrates CI’s cross-domain validity by applying it to ESA Mission 1 satellite telemetry, transforming 3.78 GB of nested binary telemetry archives into CI-compatible datasets and analyzing 58 Target-YES channels across four subsystems. Using unsupervised anomaly detection, CI produced interpretable stability diagnostics without any domain-specific tuning. Across analyzable subsystems (1, 5, 6), CI detected extremely low collapse exposure (CI = 0.000–0.006), 100% Type I Collapse patterns, zero label flips, and strong separation metrics (AUC(CI)=0.669–0.754; AUC(Confidence)=0.891–0.994). Subsystem 3 exhibited perfect prediction consistency (100% accuracy), preventing ROC calculation but confirming exceptional stability. This study provides the first cross-domain validation of the Collapse Index framework on real-world aerospace telemetry, showing that CI generalizes from artificial intelligence evaluation to natural physical systems. Results also align with CI’s broader applicability demonstrated in prior work (AI model evaluation, astrophysics, supernova detection). All analysis was performed on consumer-grade hardware (Lenovo IdeaPad 3), demonstrating CI’s computational accessibility and suitability for small-scale research environments. This upload includes the full paper, reproducibility notes, and metadata needed to support academic use.