Development

Run LAMMPS molecular dynamics with DeePMD-kit machine learning potentials. Use when the user wants MD simulations driven by a trained DP model.

Automatically logs new discoveries to CLAUDE.md files. Triggers when fixing bugs, discovering pitfalls, finding performance issues, or learning new patterns in the codebase.

Use when the user asks about HER (hydrogen evolution reaction), hydrogen adsorption free energy, or volcano plot descriptor for HER catalysts.

Generate and manage GPAW Python-based DFT calculations. Use when the user requests GPAW, Python DFT, real-space grid DFT, or LCAO-DFT with ASE integration.

Use when the user asks for Gibbs free energy, zero-point energy (ZPE), thermal corrections, or thermodynamic properties from DFT + frequency data.

Generate and manage Gaussian calculations. Use when the user requests Gaussian, G16, GJF files, or needs hybrid functionals (B3LYP), MP2, CCSD(T), or molecular quantum chemistry with Gaussian basis sets.

Use when the user asks to generate a reaction energy diagram, free energy profile, potential energy surface plot, or pathway comparison diagram for catalysis or reaction mechanism studies.

Convert between computational chemistry data formats using dpdata. Handles VASP, QE, CP2K, Gaussian, LAMMPS, and DeePMD formats. Essential for preparing ML potential training data.

Use when the user asks about density of states (DOS), projected DOS (PDOS), d-band center, spin-resolved DOS, or electronic structure analysis from completed DFT calculations.

Generate and manage DFTB+ calculations. Use when the user requests DFTB+, tight-binding DFT, SCC-DFTB, or needs fast approximate DFT for large systems or MD.

Run DeePMD-kit inference to predict energies, forces, and stresses using a trained DP model. Also covers model evaluation and testing.

CP2K single point energy calculation. Uses GTH pseudopotentials, Gaussian-plane-wave (GPW) method, and DZVP basis sets for periodic DFT.

CP2K geometry optimization. Handles bulk, slab, and molecular systems with GPW method. Efficient for large systems (200+ atoms).

Use when the user asks to test ENCUT convergence, k-point convergence, or any parameter sweep to determine converged computational settings.

Use when the user asks about CO2 reduction reaction (CO2RR), CO2 electroreduction intermediates, Faradaic efficiency, or selectivity toward CO, methanol, methane, formic acid, etc.

Use when the user asks about Bader charge analysis, charge transfer, oxidation states from DFT, or electron density partitioning.

Use when the user asks for adsorption energy, binding energy, or wants to compare how strongly a molecule binds to a surface.

Generate and manage ABINIT DFT calculations. Use when the user requests ABINIT, or needs DFPT phonons, GW calculations, or BSE optical spectra.

Recover from stale RFDB nodes that survive `commit_batch` cleanup despite having their `file` field listed in `changed_files`. Use when: (1) re-running an enricher/orchestrator step but old data with the same file field stays in the graph, (2) `commit_batch(changed_files=[X], ...)` reports success but a query later still returns nodes with `file = X`, (3) you re-pointed structural nodes (e.g., DIRECTORY/FILE) from a synthetic file path to real file paths and the old synthetic-path nodes won't...

Fix silent failures when parsing Grafema semantic IDs that are in URI format instead of legacy arrow format. Use when: (1) code splits semantic IDs by "->" but gets the whole string back because IDs are grafema:// URIs, (2) file path extraction from semantic IDs returns empty or wrong values, (3) derived edges (DEPENDS_ON, etc.) produce 0 results despite source edges existing, (4) any code that processes semantic IDs after the analysis pipeline's to_uri_format() has run. The grafema:// URI fo...

Fix node query issues in Grafema tests when nodes have numeric IDs instead of human-readable IDs, or when `type` field is undefined. Use when: (1) queryNodes returns nodes with numeric IDs like "52710336597754872375318185843222727675" instead of semantic IDs like "net:request#__network__", (2) node.type is undefined but node.nodeType has a value, (3) metadata is a JSON string instead of parsed object, (4) tests fail with "Cannot read property 'length' of undefined" after queryNodes. Covers RF...

How to write a Grafema batch-mode plugin that reads the graph and filesystem, then writes new edges/nodes directly to RFDB. Use when: (1) need to add project-specific edges that generic analyzers can't produce, (2) need to read config files (JSON/YAML) and trace values into code, (3) implementing framework-specific resolvers (Django settings, Express routes, pipeline configs). Covers: grafema.config.yaml plugin config, RFDB field naming (src/dst NOT source/target), addEdges/addNodes API, RFDB...

Cyclical dogfooding loop: test AI-AGENT-STORIES.md against the live graph, discover new stories, analyze gaps, fix root causes, re-test, write report. Use when: (1) user says "/gap-loop", (2) periodic dogfooding session, (3) after a release or major feature to re-validate stories, (4) before sprint planning to prioritize product gaps.

Systematic methodology for achieving 100% backward dataflow reachability in a new language. Create gauntlet fixture, write trace, diagnose gaps, fix analyzer/algorithm, iterate to 100%. Language-agnostic process. Use when: (1) adding a new language to Grafema, (2) auditing dataflow coverage for existing language, (3) user says "/dataflow-gauntlet".