Mobile power profiling

PATHS: resolve resources from this activated package, not the app repository. This skill supplies capture/analysis procedures and a descriptive comparison CLI, not installed profilers, device access or fabricated power data.

Step 1: Define the question and measurement boundary

1. RECORD the symptom or implementation comparison, app/build, physical device/model/OS, tool version, workload, foreground/background state and evidence already available. Separate the capture device from the analysis host.
2. iOS/iPadOS -> READ references/ios.md. Android -> READ references/android.md. For a cross-platform comparison, run separate experiments; platform scores and sensors are not one common energy scale.
3. RECORD metric, unit and attribution: Apple overall system battery-energy fraction/hour versus relative app subsystem impact; Android ODPM whole-device rail power versus battery-gauge observations. Select app/process explicitly where the tool supports app impact.
4. READ references/sources.md when reconciling the five source articles, old UI instructions, version eligibility or broad battery-saving claims. Primary platform documentation governs; numeric demonstrations are not universal thresholds.
5. Obtain authorization before Developer Mode/debugging changes, elevated/device capture, app restarts or sharing traces. Preserve source recordings and sensitive identifiers; perform local analysis without unapproved uploads.

DONE: the supported capture route, required signal, units, device/app identity and attribution limits are explicit. No eligible tool/device -> name that prerequisite and complete reachable analysis, not simulated measurement.

Step 2: Plan and capture comparable evidence