torch-geometric

PyTorch Geometric (PyG) for graph neural networks — node/link/graph classification, message passing (GCN, GAT, GraphSAGE, GIN), heterogeneous graphs, neighbor sampling, and custom datasets. Use when working with torch_geometric, not for general NetworkX analytics or non-graph PyTorch models.

Quality

89%

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

86%Weight 40%Scale 1-5

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

A high-quality, code-dense skill body that is executable, well-structured, and effectively offloads depth to six clearly signaled reference files. Minor tightening of explanatory prose and a few more validation checkpoints would push it to full marks.

Suggestions

Tighten the prose around the conv-layer selection table and the _i/_j indexing convention to rely more on inline code comments.

Add an explicit validation/checkpoint note for the NeighborLoader and link-prediction workflows (e.g., verify seed-node slicing against expected loss shape).

Consider moving the PyG 2.7 version-compatibility notes into the scaling or a dedicated compatibility reference to keep the installation section leaner.

Dimension	Reasoning	Score
Conciseness	Largely lean and code-forward, assuming Claude's competence (no padding about what graphs/libraries are); a few prose passages like the conv-layer table and _i/_j convention explanation could be trimmed slightly.	4 / 5
Actionability	Dense, copy-paste-ready executable code covering the common cases — Data construction, GCN training, graph classification with pooling, link prediction splits, NeighborLoader, and HeteroData — with concrete commands and imports.	5 / 5
Workflow Clarity	Multi-step workflows (install order, training/eval loops, NeighborLoader seed-node slicing) are clearly sequenced with a pitfalls checklist, though destructive/batch operations needing validate-retry feedback loops are minimal here.	4 / 5
Progressive Disclosure	SKILL.md serves as a clear overview with well-signaled, one-level-deep pointers ('Read references/message_passing.md ...') to six real, topically-split reference files, keeping the body navigable.	5 / 5
	Total	18 / 20 Passed

Description

92%Weight 40%Scale 1-5

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 names concrete capabilities, provides explicit positive and negative triggers, and clearly distinguishes the skill from adjacent graph/ML libraries. Only minor keyword-variation gaps keep trigger quality just below full marks.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capabilities — 'node/link/graph classification, message passing (GCN, GAT, GraphSAGE, GIN), heterogeneous graphs, neighbor sampling, and custom datasets' — giving comprehensive coverage of the library's surface.	5 / 5
Completeness	Explicitly answers both what (the listed capabilities) and when ('Use when working with torch_geometric'), plus a concrete negative boundary ('not for general NetworkX analytics or non-graph PyTorch models').	5 / 5
Trigger Term Quality	Strong natural keyword coverage including synonyms ('PyTorch Geometric (PyG)', 'graph neural networks', 'torch_geometric', 'NetworkX', 'PyTorch models'), though a few common user phrasings are absent.	4 / 5
Distinctiveness Conflict Risk	Clear niche (PyG for GNNs) with explicit disambiguation against NetworkX and general PyTorch, minimizing trigger overlap with neighboring skills.	5 / 5
	Total	19 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: K-Dense-AI/scientific-agent-skills
Path: skills/torch-geometric/SKILL.md
Commit: b2a92ba

Reviewed: 3 days 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.