Tools¶

This page describes implemented Python tools and agent wrappers in src/cspilot/tools/.

ASE¶

Module: ase_tools.py

Function	Purpose
`load_structure`	Read a structure with ASE.
`save_structure`	Write a structure with ASE.
`summarize_structure`	Return formula, atom count, cell, PBC, and center of mass.

Agent wrapper: inspect_structure.

Molecule Conversion¶

Module: mol_tools.py

Function	Purpose
`molecule_name_to_smiles`	PubChem name lookup.
`validate_smiles`	RDKit SMILES validation.
`canonicalize_smiles`	RDKit canonical SMILES.
`smiles_to_xyz`	RDKit ETKDG conformer generation and MMFF/UFF minimization.
`molecule_name_to_xyz`	PubChem name lookup followed by RDKit XYZ generation.

Agent wrappers include molecule_name_to_smiles_tool, smiles_to_xyz_tool, molecule_name_to_xyz_tool, validate_smiles_tool, and canonicalize_smiles_tool.

stk¶

Module: stk_tools.py

Function	Purpose
`stk_build_from_smiles`	Build an stk/RDKit molecule from SMILES and write `.mol`, `.sdf`, or `.xyz`.
`stk_building_block_from_file`	Load an stk building block from a molecule file.
`stk_linear_polymer_from_smiles`	Construct a linear polymer with a whitelisted stable path.
`stk_construct_cage_from_smiles`	Returns a clear unsupported result for cage topologies until safe presets are implemented.
`rdkit_replace_substructure`	RDKit-based molecule editing.
`stk_export_to_xyz`	Export molecule files to XYZ with stk/RDKit/ASE fallback.

CLI commands: stk-build-smiles, stk-polymer, stk-xtb, and the stk subcommands.

xTB¶

Module: xtb_tools.py

Function	Purpose
`optimize_with_xtb`	Run xTB geometry optimization through the configured command.

CLI: xtb-opt. Agent wrapper: run_xtb_optimize.

ORCA / OPI¶

Module: opi_orca_tools.py

Function	Purpose
`orca_single_point`	OPI ORCA single point.
`orca_optimize`	OPI ORCA optimization.
`orca_frequency`	OPI ORCA frequency.
`parse_orca_result`	OPI parsing plus fallback text parsing.
`single_point_with_orca`	CLI adapter for `orca-sp`.

Parsed values may include energies, orbital data, dipole, thermochemistry, and frequencies when those values exist in ORCA output. Missing values are not invented.

MACE¶

Module: mace_tools.py

Function	Purpose
`optimize_with_mace`	ASE optimization with `MACECalculator`.

CLI: mace-opt. Agent wrapper: run_mace_optimize.

AGAPI Materials¶

Module: agapi_materials_tools.py

Function	Purpose
`agapi_materials_query`	Optional wrapper around AGAPI prebuilt materials query.

Agent wrapper: agapi_materials_query_tool.

Result Tools¶

Module: result_tools.py

Function	Purpose
`find_result_json`	Recursively find result JSON files.
`load_result_json`	Load a JSON result file.
`get_property_from_result`	Find a property through recursive lookup and aliases.

Agent wrappers: find_result_json_tool, get_property_from_result_tool, and name-overridden direct-agent variants.

NWPESSe¶

Module: nwpesse_tools.py

Implemented helpers parse fragment formulas, generate validated box configs, write mol.cluster and mol.inp, run the external NWPESSe binary, parse candidate XYZ second-line energies, and copy the lowest-energy structure.

CLI: nwpesse-search. Agent wrapper: nwpesse_global_minimum_search_tool.

GreenCatAI¶

Module: greencatai_tools.py

greencatai_design_mbh_catalysts calls the installed GreenCatAI MBH design API when available.

CLI: greencatai design-mbh. Agent wrapper: design_mbh_catalysts_tool.