What problem does it solve? Running the SFINCS flood model correctly requires deep operational knowledge: binary input formats, unit conversions, vertical datum alignment, CFL-stable timesteps, and boundary mask conventions. Mistakes in any of these produce silent failures where the model runs but outputs wrong flood depths. This Skill packages that expertise so an AI agent can set up, execute, and validate SFINCS simulations without silent errors. ## Core Features & Use Cases - End-to-end 8-stage pipeline: 13 validated Python tools covering domain setup, DEM/topography processing, Manning's roughness, rainfall and CaMa-Flood boundary forcing, configuration generation, execution, and flood map post-processing. - 44 diagnostic triplets: A symptom-to-remedy database covering unit conversion traps (mm/3hr vs mm/hr), binary format pitfalls (sfincs.ind), mask boundary errors, and coupling double-counting, so failures are diagnosed instead of guessed. - Model coupling: Converts CaMa-Flood discharge/water level and CMFD/MSWX precipitation into SFINCS boundary conditions for compound coastal-fluvial-pluvial flood modeling. - Use Case: Simulate the July 2003 Huai River flood at 100m resolution by chaining VIC runoff, CaMa-Flood routing, and SFINCS inundation, then validate flood extent against satellite observations using CSI scoring. ## Quick Start Ask the agent to run the preflight check and then execute the SFINCS pipeline for your basin shapefile, from domain setup through flood map generation.