candidate-generator

Generate inorganic crystal structure candidates for DFT screening and ML datasets.

7|1|Updated Mar 13, 2026
One-click install
npx skills add https://github.com/hkqai/MatClaw --skill candidate-generator-hkqai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: candidate-generator
Source: https://github.com/hkqai/MatClaw/tree/main/skills/candidate-generator
Command: npx skills add https://github.com/hkqai/MatClaw --skill candidate-generator-hkqai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill eliminates the manual and ad hoc process of producing diverse, physically plausible inorganic crystal structure candidates for high-throughput DFT screening, machine learning dataset construction, and materials discovery campaigns.

Core Features & Use Cases

  • End-to-end candidate pipeline: composition enumeration, prototype building, chemical substitution and ion exchange, disorder resolution (enumeration or SQS), defect generation, and structural perturbation/augmentation.
  • Integrated filtering and routing: charge-neutrality checks, Ewald ranking, Materials Project cross-checks, and ASE-format output for direct database storage and DFT workflows.
  • Use Cases: discover Li-Mn-P-O cathode candidates from elements-only input, generate isostructural analogues via ICSD-informed substitution, create SQS for high-entropy oxides, and produce defect supercells for targeted defect engineering studies.

Quick Start

Generate a diverse set of Li-Mn-P-O candidate structures from elements-only input, resolve disorder and defects as needed, and save ASE-formatted results to the candidates database for downstream DFT screening.

Frequently Asked Questions about candidate-generator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate inorganic crystal structure candidates from elements-only input for DFT screening?▼

Generate inorganic crystal structure candidates by inputting desired elements to trigger composition enumeration, prototype seeding, and chemical substitution, yielding ASE-formatted structures ready for DFT screening and database ingestion.

Can I create special quasirandom structures for high-entropy oxides using pymatgen and ASE?▼

Create special quasirandom structures (SQS) for high-entropy oxides using the skill's disorder resolution methods, which leverage pymatgen-based enumeration tools and output ASE-formatted results for downstream analysis.

What is the best way to generate charge-neutral defect supercells for materials discovery workflows?▼

Generate charge-neutral defect supercells by using the defect generation feature, which applies charge-neutrality checks and Ewald ranking to produce physically plausible structures for targeted defect engineering studies.

How does chemical substitution and ion exchange work for generating isostructural analogues?▼

Chemical substitution and ion exchange generate isostructural analogues by applying charge-neutral ion exchange and ICSD-informed pymatgen substitution tools to existing prototype structures, producing diverse new candidates.

Does this approach support ASE database storage for machine learning dataset construction?▼

ASE database storage is fully supported, allowing you to save generated candidates directly in ASE-format for seamless machine learning dataset construction and integration into high-throughput computational workflows.

What filtering and ranking methods are applied when enumerating compositions and resolving disorder?▼

Filtering and ranking during composition enumeration and disorder resolution include charge-neutrality checks, Ewald ranking, and Materials Project cross-checks to ensure only physically plausible structures proceed to DFT screening.