hdf5-pde-data-loading

Patterns for loading PDE simulation datasets (PDEBench, PhiFlow, JAX-CFD) from HDF5 files. Handles layout detection (single tensor vs separate variables), spatial/temporal downsampling, multi-variable systems, HuggingFace and DaRUS data sources, and efficient PyTorch DataLoader creation. Use when preparing PDE data for neural operator training.

Quality

93%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

87%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A lean, highly actionable single-file skill with executable code and compact high-signal tables. The main gap is the absence of explicit validation checkpoints in the load workflow, which the rubric penalizes for batch/risky operations.

Suggestions

Add an explicit verification checkpoint after loading — e.g., assert the expected [N, X, T, C] shape and dtype before returning data, and re-download on size mismatch — to close the workflow-validation gap.

Insert a short validation step in the DaRUS download flow (check file size / md5 against the CSV after aria2c) so partial downloads are caught as a checkpoint rather than only handled reactively in the pitfalls table.

Consider splitting the full load_pde_hdf5 implementation or DaRUS sourcing detail into a references/ file if the skill grows, keeping SKILL.md as an overview pointing one level deep.

Dimension	Reasoning	Score
Conciseness	The body is dense and high-signal: executable code, compact tables, and short notes with no explanation of concepts Claude already knows (e.g., no "what is HDF5"). Every section earns its place; it does not fall to the score-2 'could be tightened' band.	3 / 3
Actionability	Provides fully executable, copy-paste-ready code (load_pde_hdf5 loader, hf_hub_download with concrete repo_id/filename, aria2c download command, PDEDS Dataset and DataLoader) plus specific downsample factors — matching the score-3 anchor for concrete, executable guidance.	3 / 3
Workflow Clarity	Sections are logically sequenced (detect layout → load → source data → build DataLoader → downsample) and the Common Pitfalls table gives error→fix recovery guidance, but the load workflow lacks explicit validation checkpoints (e.g., verify file integrity / shape before proceeding), which the rubric says should cap batch/risky operations at 2.	2 / 3
Progressive Disclosure	No bundle files exist and the skill is a single, well-organized file with clear sections (When to Use, Layout Detection, Data Sources, DataLoader, Downsampling, Pitfalls) and no nested references; this self-contained organization satisfies the rubric's allowance for skills with no external-reference needs.	3 / 3
	Total	11 / 12 Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that uses third-person voice, names concrete capabilities, and includes an explicit 'Use when' trigger for a well-defined niche. It would not benefit from padding or vague qualifiers.

Dimension	Reasoning	Score
Specificity	Names the domain and multiple concrete actions: "layout detection (single tensor vs separate variables), spatial/temporal downsampling, multi-variable systems, HuggingFace and DaRUS data sources, and efficient PyTorch DataLoader creation" — a comprehensive list of specific capabilities matching the score-3 anchor.	3 / 3
Completeness	Clearly states what it does (loading, layout detection, downsampling, multi-variable, data sources, DataLoader) and provides an explicit trigger ("Use when preparing PDE data for neural operator training."), satisfying both what and when.	3 / 3
Trigger Term Quality	Includes the natural terms a user would say for this niche — "HDF5", "PDE", "PDEBench", "PhiFlow", "JAX-CFD", "DataLoader", "neural operator training" — giving good coverage rather than generic or purely jargon phrasing.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear niche (HDF5 PDE simulation data for neural operator training) with domain-specific triggers unlikely to fire for unrelated skills.	3 / 3
	Total	12 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/data-engineering/hdf5-pde-data-loading/SKILL.md
Commit: 52845c3

Reviewed: about 11 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.