Claude Skills by lzwei196

> **Stage ID**: s4_agromanagement > **Pipeline order**: 4 of 8 > **Depends on**: s1_crop_params (crop_name and variety_name must match)

> **Stage ID**: s5_engine_config > **Pipeline order**: 5 of 8 > **Depends on**: s1_crop_params, s2_soil_params, s3_weather_prep, s4_agromanagement

> **Stage ID**: s6_execution > **Pipeline order**: 6 of 8 > **Depends on**: s5_engine_config

> **Stage ID**: s7_output_parsing > **Pipeline order**: 7 of 8 > **Depends on**: s6_execution

> **Stage ID**: s8_yield_analysis > **Pipeline order**: 8 of 8 > **Depends on**: s7_output_parsing

> **Version**: 4.7.1 > **Domain**: Mesoscale atmospheric simulation > **Last updated**: 2026-08-18 > **Validation status**: partial_replacement ---

**Package**: `hydrocraft-wrfhydro-standalone` v2.2.0 **Model**: WRF-Hydro v5.2.0 offline (NoahMP land surface + gridded routing) **Created by**: Jianyun Zhang Research Group, Hohai University **Last updated**: 2026-03-21 (4 routing skill documents added) **Stats**: 14 tools | 6 skill documents | 46 diagnostic triplets | 28 error log entries | ~6,200 lines of validated Python ---

> **Stage ID**: s11_output_interpretation > **Pipeline order**: 11 of 12 (post-execution analysis) > **Depends on**: s10_execution (completed WRF-Hydro run)

> **Stage ID**: s4_channel_routing > **Pipeline order**: 4 of 12 (part of Fulldom construction + namelist configuration) > **Depends on**: s1_domain, s2_geo_em, s4_fulldom, s9_namelists

> **Stage ID**: s6_groundwater > **Pipeline order**: 6 of 12 (part of groundwater/ancillary file construction) > **Depends on**: s1_domain, s2_geo_em

> **Stage ID**: s9_runoff_options > **Pipeline order**: 9 of 12 (configured in namelist.hrldas, Stage s9_namelists) > **Depends on**: s5_soil_properties, s9_namelists

- **Package**: `hydrocraft-wsimod` - **Model**: WSIMOD (Water Systems Integrated Modelling framework) - **Version**: 0.6+ (PyPI) - **Domain**: Integrated water cycle modelling — water quantity and quality - **Authors**: Barnaby Dobson, Imperial College London - **References**: Dobson et al. 2023 (JOSS, doi:10.21105/joss.04996); Dobson et al. 2024 (GMD, doi:10.5194/gmd-17-4495-2024) - **Stats**: 4 tools, 5 skill docs, 18 diagnostic triplets, ~11,500 lines Python source ---

Configure the WSIMOD simulation: define the study period, select pollutants to track, design the node-arc graph topology, and choose execution mode (CLI vs Python API). This stage produces the structural blueprint that all downstream stages depend on.

Convert raw meteorological and hydrological timeseries data into WSIMOD's `data_input_dict` format. This is the most error-prone stage because unit conversions are silent — wrong units produce plausible-looking but wrong results.

Set physical parameters for each WSIMOD node: surface areas, storage capacities, residence times, pollutant loads, treatment efficiencies, and soil properties. Incorrect parameters produce silent errors because the model runs regardless.

Run the WSIMOD simulation either through the CLI entry point or the Python API. Understand the orchestration sequence, handle runtime errors, and capture output.

Parse, validate, and analyze WSIMOD output files (flows.csv, tanks.csv, surfaces.csv). Extract specific variables, compute water balance metrics, and generate visualizations for model validation and reporting.

**Package**: `hydrocraft-dfnworks` v1.0.0 **Model**: dfnWorks v2.10.0 (DFNGen v2.3, DFNTrans, pydfnworks) **Created by**: LANL EES-16 (Jeffrey Hyman, Daniel Livingston, Satish Karra) **Last updated**: 2026-03-25 **Stats**: 4 tools | 6 skill documents | 17 diagnostic triplets | ~2,500 lines of validated Python **Validation status**: `development_validated` (graph-based flow/transport on 3-family TPL network) ---

Configure the 3D domain geometry, define fracture families with statistical distributions, set boundary conditions, and assign hydraulic properties. This stage produces the complete parameter set required by DFNGen (Stage 1). All subsequent stages depend on correct configuration.

Generate a 3D discrete fracture network (DFN) using the DFNGen C++ executable. Places fractures stochastically in the domain according to specified statistical distributions for size, orientation, and spatial density. Rejects fractures that violate geometric constraints (too close to existing fractures, outside domain, etc.).

Generate a conforming Delaunay triangulation of the discrete fracture network using LaGriT. The mesh resolves all fracture intersections as triangle edges and creates Voronoi control volumes suitable for finite volume flow solvers (PFLOTRAN, FEHM). This stage is required for full-physics flow simulation but is NOT needed for graph-based mode.

Solve steady-state fluid flow through the discrete fracture network to obtain pressure and velocity fields. Three solvers are available:

Simulate solute transport through the DFN using Lagrangian particle tracking. Particles are released at the inlet boundary and tracked through the fracture network to the outlet, producing breakthrough curves (travel time distributions) that characterize transport behavior.

Extract, analyze, and visualize dfnWorks simulation results. Compute summary statistics, breakthrough curves, effective permeability, flow channeling metrics, and generate publication-quality figures.

Convert discrete fracture network (DFN) results to equivalent continuum representations for coupling with larger-scale models. Supports upscaled equivalent porous medium (ECPM) and unstructured DFM-to-continuum mapping for integration with basin-scale groundwater models.

**Package**: `hydrocraft-icepack-glacier` v1.0.0 **Model**: icepack v1.1.0 — Glacier Flow Modeling with Finite Elements **Author**: Daniel Shapero, University of Washington **Domain**: Cryosphere (ice sheets, ice shelves, ice streams, glaciers) **Last updated**: 2026-03-26 **Stats**: 4 tools | 6 skill documents | 20 diagnostic triplets **Validation status**: `structurally_validated` ---

> **MANDATORY EXECUTION POLICY** — READ BEFORE PROCEEDING > > You MUST run the **actual model binary or package** described in this document. > If the model fails to import, compile, or execute, you MUST: > 1. Check `diagnostics/triplets.yaml` for a matching error pattern > 2. Try the fix suggested in the triplet's `remedy` section > 3. If still failing, report the error to the user with full details > > You MUST NOT substitute a simplified Python formula, regression equation, > or hand-coded...

Create the mHM run directory tree and `config.json` for one basin. This stage fixes the basin name, period, three-grid resolutions, forcing source, PET mode, and mHM binary path used by all later tools.

Transfer calibrated MPR global parameters from a gauged source basin to a target basin. This is the main mHM regionalization workflow in this KI.

Define the true catchment and the three mHM grids: L0 morphology, L1 hydrology, and L11 routing. This KI treats MERIT-Hydro area snapping as the first domain action for flat or engineered basins.

Generate every L0 morphological input mHM needs as aligned ESRI ASCII grids and matching classdefinition files: terrain, flow direction/accumulation, soil, geology, land cover/LAI, and gauge grid.

Create `mhm_parameter.nml`, the global MPR transfer-function parameter file. These are process parameters, not per-cell calibrated maps, and later s9/s10 stages modify or transfer them.

Convert gridded meteorological forcing into the NetCDF layout and units mHM reads: `pre`, `tavg`, and, for PET method 0, `pet`.

Convert observed discharge to the text format mHM reads for gauge evaluation.

Assemble the four Fortran namelists mHM needs, including grid settings, process switches, paths, gauge metadata, evaluation period, optional optimization settings, and the parameter file.

Run the actual mHM binary for the prepared basin. This stage must execute `KISSPATH_BINARIES/mhm/mhm` or the binary path in the run config; do not replace it with a simplified formula.

Extract and evaluate mHM outputs, especially routed discharge, actual ET, soil moisture, and cross-model discharge comparisons.

Configure and run mHM's built-in DDS or SCE optimizer, then extract the best MPR global parameters for later forward validation or s10 transfer.

**Package**: `hydrocraft-mizuroute` v1.0.0 **Model**: mizuRoute (NCAR/ESCOMP), reach-based river routing **Repository**: https://github.com/ESCOMP/mizuRoute **Created by**: Jianyun Zhang Research Group, Hohai University **Last updated**: 2026-03-22 **Stats**: 7 tools | 8 skill documents | 20 diagnostic triplets | 2,115 lines of validated Python **Validation status**: `production_validated` — Bengbu IRF routing, r=0.87 vs Lohmann, mass-conserving (Q=1795 m3/s), 3 serial stub bugs fixed ---

Select the appropriate routing method, configure basin parameters, and determine whether lake routing is needed. This stage produces the configuration consumed by all downstream stages.

Build the river network topology NetCDF that defines every reach and its contributing catchment (HRU). This is the core spatial input for mizuRoute.

Create area-weighted spatial mapping from VIC regular grid cells to mizuRoute HRU catchments. Each VIC cell's runoff is distributed proportionally to the HRUs it overlaps.

Convert VIC or WRF-Hydro runoff output to mizuRoute input NetCDF format with correct units. **This is the #1 silent error risk in VIC-mizuRoute coupling.** Getting the unit conversion wrong produces discharge that is 8x, 24x, or 86400x too high or low with NO error message.

Generate the mizuRoute control file that specifies all input/output paths, variable names, routing method, and time settings.

Execute the mizuRoute Fortran binary with proper validation before and after.

Extract discharge from mizuRoute output, compare with observations and other routing models (Lohmann, CaMa-Flood), and evaluate the 5 routing methods against each other.

Configure mizuRoute to route flow through lakes and reservoirs using one of three available schemes. Required for basins with significant storage (reservoirs, natural lakes).

**Package**: hydrocraft-openamundsen-snow **Version**: 1.0.0 **Created**: 2026-03-26 **Model**: openAMUNDSEN (fully distributed snow/hydroclimatological model) **Domain**: Cryosphere — snow accumulation, ablation, energy balance in mountain regions **Language**: Python (with Numba JIT acceleration) **License**: MIT ---

**Package**: `hydrocraft-badlands-geomorph` v1.0.0 **Model**: pyBadlands (Basin and Landscape Dynamics) — Tristan Salles, University of Sydney **Domain**: Geomorphology / Landscape Evolution Modelling **License**: GNU LGPL v3 **Language**: Python + Fortran + C extensions **Build**: Meson + mesonpy **Stats**: 7 pipeline stages | 4 tools | 116 source files | 20+ diagnostic triplets ---

**Package**: hydrocraft-pygimli-geophysics v1.0.0 **Model**: pyGIMLi (Python Library for Geophysical Inversion and Modelling) **Version**: 1.5+ **Domain**: Geophysics — Electrical Resistivity Tomography (ERT), Seismic Refraction (SRT), Electromagnetics (EM), Induced Polarization (IP) **Authors**: Carsten Rücker, Thomas Günther, Florian Wagner **License**: Apache 2.0 **Stats**: 4 tools | 5 skill documents | 18 diagnostic triplets | ~71,000 lines Python ---