القائمة
Use when running ELFQuake prospective real-data refresh scripts, including relabeling matured VLF rows and rebuilding real model input artifacts.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | elfquake-data-refresh |
| description | Use when running ELFQuake prospective real-data refresh scripts, including relabeling matured VLF rows and rebuilding real model input artifacts. |
Run from the repository root. Keep data Italy-scoped, preserve raw records, normalize timestamps to UTC, and do not claim prediction value from refreshed data alone.
./scripts/refresh-prospective-labels.sh
./scripts/prepare-real-model-inputs.sh
rg -n "status|labeled_row_count|positive_count|negative_count" data/derived/models/*real_vlf_aligned_windows.readiness.json data/derived/multimodal/*labeled.readiness.json
Only run real training when readiness shows both positive and negative examples. If class variation is absent, report the blocker and keep collecting data.
Use these only when relevant:
./scripts/materialize-real-vlf-sequence.sh
./scripts/train-real-tabular-model.sh
After meaningful count or status changes, update docs/report.md and docs/next-actions.md concisely.
Use when running or regenerating ELFQuake sandpile simulations, synthetic seismic events, piezo/VLF-like signals, maps, videos, or simulation-derived model artifacts.
Use when running one-off ELFQuake source fetch, capture, normalization, or feature-build CLI commands for INGV, VLF, astronomy, geomagnetic, or space-weather data.
Use when training, comparing, diagnosing, or selecting ELFQuake synthetic PyTorch models, including multimodal sequence models and missing-modality ablation checks.