Complex interactive artifacts require more structure than a single HTML file. This skill scaffolds a React project and bundles it into shareable HTML.

Transformer workflows require careful choices for models, pipelines, tokenizers, generation, and training. This skill provides focused Hugging Face patterns for implementation and optimization.

Symbolic mathematics becomes difficult when exact forms, assumptions, and solver choices matter. This skill provides focused SymPy workflows for reliable exact computation.

Reinforcement learning projects require careful algorithm, environment, training, and evaluation choices. This skill provides Stable Baselines3 guidance and reusable templates for each stage.

Frontend work often needs consistent React patterns, performance checks, and project scaffolding. This skill gives Claude and compatible agents practical guidance and local helper scripts for modern web application development.

Computer vision projects often fail when model choices ignore data quality, latency, deployment, or monitoring constraints. This skill structures architecture, optimization, and production planning around those constraints.

Architecture choices are hard to compare across product, scale, security, and team constraints. This skill gives structured guidance, references, and helper scripts for clearer decisions.

Quantum computing projects require coordinated circuit design, transpilation, execution, and result analysis. This skill provides practical Qiskit guidance for simulations, algorithms, visualizations, and supported quantum hardware.

Deep learning projects become difficult to maintain when training, data, logging, and scaling logic are mixed. This skill provides structured PyTorch Lightning guidance and templates.

Reinforcement learning projects often lose time on environment integration, policy design, and slow simulation. This skill provides focused PufferLib templates and optimization guidance.

Quantum workflows require careful circuit, device, gradient, and framework choices. This skill provides PennyLane patterns for building, training, and evaluating those workflows.

Designing reliable robot protocols requires precise API, labware, pipette, and module details. This skill provides focused guidance and templates for Opentrons Flex and OT-2 workflows.

Move projects can drift from modern syntax and package conventions. This skill reviews Move packages against the Move Book checklist and returns prioritized, actionable findings.

Pull request failures and scattered review feedback slow delivery. This skill investigates evidence, applies targeted fixes, and repeats checks until the branch is ready.

Vague commits make reviews and release notes harder. This skill analyzes staged Git changes and drafts structured Conventional Commit messages.

Review feedback can be difficult to organize and resolve. This skill summarizes open PR comments, asks which items to address, and guides selected fixes.

Debugging errors wastes time when messages lack context. This skill guides classification, root cause analysis, fixes, verification, and reusable solution records.

Sequential investigation slows work when failures have unrelated causes. This skill helps Claude, Codex, and Claude Code delegate focused tasks and verify combined results.

Quantum workflows require careful circuit design, simulation, optimization, and provider setup. This skill provides practical Cirq guidance for simulators, noise studies, experiments, and hardware execution.

Early feature ideas often hide unclear goals, constraints, and technical risks. This skill guides focused discovery, approach comparison, and incremental design validation before implementation.