ds-lite-engineering

What problem does it solve? Engineering analyses often fail silently through wrong units, incorrect FFT scaling, aliasing, leakage, or invented physical parameters. This Skill enforces a bounded, auditable workflow for numerical and signal-processing work so every analysis declares its units, sampling, seeds, and checks before execution. ## Core Features & Use Cases - Backend discovery: Detects whether Python/NumPy/SciPy, MATLAB, or Octave is actually installed before running, and never installs or assumes unavailable tools. - FFT and signal validation: Checks frequency resolution, Nyquist limits, windowing leakage, amplitude/power scaling, and figure-axis units. - Evidence-linked results: Validates output against the ds-lite.engineering-analysis.v1 schema and links results to a Core Evidence Pack with unverified items and a next action. - Use Case: You need to verify an FFT of a sampled vibration signal. The Skill requires you to state sampling rate, duration, window, and scaling, runs the analysis in Python, and flags aliasing or axis-unit errors before the result is recorded. ## Quick Start Use ds-lite-engineering to define and audit one bounded FFT analysis of my sampled time series with explicit units, window, scaling, and seed.