What problem does it solve? Working with CAE and scientific datasets in Kit-CAE requires knowing the correct launch flags, import APIs, stage discovery patterns, and Z-up camera conventions, which are easy to get wrong and cause GPU crashes or stale renders. This Skill centralizes that shared foundation so dependent CAE skills and agents do not re-document or misapply the same primitives. ## Core Features & Use Cases - Preflight and Launch Guidance: Checklists for build validation, GPU requirements, shader cache warmup, and mandatory launch arguments that prevent ERROR_DEVICE_LOST crashes during rendering. - Unified Import and Stage Discovery: Documents the import_to_stage dispatcher for CGNS, VTK, EnSight, OpenFOAM, NPZ, and other formats, plus OmniSci dataset and field discovery patterns. - Ready-to-Run Inspection Scripts: Bundled scripts inspect CGNS/VTK files, query field statistics as JSON, and warm up the Kit shader cache, all executed through Kit's --exec environment. - Use Case: An agent needs to inspect a CGNS simulation file before visualizing it. It runs the inspect_cgns.py script via repo.sh launch, receives a JSON listing of datasets, fields, and associations, then uses the API reference to create visualization operators with correct field selections. ## Quick Start Ask the agent to inspect a CGNS or VTK file by running the bundled inspection script inside Kit-CAE with the CAE_INSPECT_FILE environment variable set to your data file.
