hpc-cluster
Run jobs on CU Boulder CURC HPC cluster (Alpine). Use when simulations need more compute than the local workstation, for large-scale parallel jobs, or when GPU resources are needed beyond local availability. You have full SSH access - work like a researcher.
campaign-orchestration
Orchestrate one or more long-running compute campaigns (HPC simulations, multi-stage MD pipelines, ensemble runs) where stages take hours to days and failures must not halt the pipeline. Use when running multiple parallel campaigns that share a stage structure (smoke → equilibration → production → analysis), when waiting on long jobs makes interactive driving impractical, or when designing a self-healing tick-based supervisor for compute work. Reads/writes per-campaign WORKFLOW.md files as durable state.
compute-strategy
Decide where and how to run compute jobs across available backends (HPC, cloud GPUs, local). Use when starting any compute-intensive task, choosing a backend or partition, deciding job parameters (walltime, GPU type, parallelism), or debugging compute infrastructure issues. Embeds the smoke-first iteration pattern — validate cheaply before scaling.
compute-validation
Validate that an expensive long-running compute campaign is ready for production. Combines reasoning-based verification (predict failure modes from system analysis + priors, before any compute) with empirical smoke analysis (run a cheap smoke, extract predictive signal, extrapolate to production behavior). Iterate until satisfied. Use before submitting any HPC simulation, ML training run, DFT calculation, or other expensive compute. Composes with compute-strategy (WHERE to run) and campaign-orchestration (HOW to manage long execution).
data-analysis
Analyze simulation data and compute properties. Use when asked to parse LAMMPS/QE output, calculate diffusion coefficients, RDF, MSD, energies, or generate plots and visualizations.
iff-parameters
Search, retrieve, compare, and export INTERFACE Force Field (IFF) parameters for molecular dynamics simulations. Use BEFORE searching literature for force field parameters — IFF covers metals, minerals, oxides, clays, and organic interfaces with validated, citation-ready values.
lammps-simulation
Run LAMMPS molecular dynamics simulations. Use when asked to run MD simulations, energy minimization, equilibration, production runs, or calculate properties like diffusion, RDF, MSD. Supports both CPU and GPU execution.
literature-search
Search and retrieve scientific literature. Use when asked to find papers, research a topic, find citations, get paper abstracts, or conduct literature reviews. Accesses Semantic Scholar, arXiv, and other academic databases.
materials-database
Query materials databases for structures and properties. Use when asked to get crystal structures, material properties, phase diagrams, or thermodynamic data. Primary source is Materials Project, with NIST, PubChem as secondary.
project-update
Get fully up to date on a project, or build its weekly / PI-meeting update bundle, from INSIDE the project directory. Use when an agent lands cold in a project and needs the current state (open workstreams, what moved, open PI decisions, blockers), or when asked to "build the weekly update", "prep the PI meeting", or "aggregate the relevant files into the update dir". Mines only THIS repo, driven by its own .sync/manifest.yaml. No dependency on LabSync.
quantum-espresso
Run Quantum ESPRESSO DFT calculations. Use when asked to perform first-principles calculations, SCF, structural relaxation, band structure, DOS, phonons, or any ab initio quantum mechanical calculation.
resource-acquisition
Find and acquire computational science resources autonomously. Use when you need force field parameters, pseudopotentials, crystal structures, or any other scientific data. You are a researcher - you find what you need.
theory-synthesis
Synthesize scientific theories from literature corpora using ASTA Theorizer. Use when asked to generate hypotheses, find research gaps, synthesize knowledge across many papers, or create literature-grounded theories for computational testing.
vast-cloud
Run jobs on VAST AI cloud GPUs. Use when you need GPU compute immediately without queue times, for short-to-medium jobs (<4 hours), or when HPC is unavailable. Pay-per-hour pricing.