Skills torch-geometric

📦

torch-geometric

Name: torch-geometric
Author: davila7

Content revision r2 Safe 📁 Filesystem access⚡ Contains scripts⚙️ External commands🌐 Network access

Build Graph Neural Networks with PyTorch Geometric

Graph learning projects require specialized data structures, layers, sampling, and evaluation workflows. This skill provides PyTorch Geometric guidance, references, templates, visualization, and benchmarking utilities.

Supports: Claude Codex Code(CC)

🥈 80 Silver

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request

Review the Skillstore skill "torch-geometric" from https://skillstore.io/skills/davila7-torch-geometric.md and its manifest at https://skillstore.io/api/skills/davila7-torch-geometric/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/davila7-torch-geometric.md Signed manifest GET /api/skills/davila7-torch-geometric/manifest Signed lockfile GET /api/skills/davila7-torch-geometric/lockfile

Test it

Using "torch-geometric". Choose a model for citation network node classification.

Expected outcome:

Recommendation: use GCN for a clear baseline, then test GAT when neighbor importance may vary.
Workflow: normalize features, use provided masks, track validation accuracy, and report the selected test result.
Checks: confirm edge direction, class balance, isolated nodes, and feature dimensions.

Using "torch-geometric". Plan a fair comparison of GCN, GAT, and GraphSAGE.

Expected outcome:

Use identical data splits, epoch budgets, seeds, hidden dimensions, and evaluation metrics.
Report accuracy and training time across repeated runs with mean and variation.
Record hardware, package versions, and memory limits so the comparison can be reproduced.

Using "torch-geometric". Outline a heterogeneous recommendation model.

Expected outcome:

Represent users and items as separate node types with typed interaction edges.
Select relation-specific message passing, define negative samples, and prevent temporal leakage.
Evaluate ranking quality by user segment and inspect explanations for influential relations.

Security Audit

Safe

v11 • 7/23/2026 Open versioned report

The audit reviewed all 175 static findings. One medium-risk issue is confirmed because the template generator can overwrite any writable path supplied through its output argument. All other alerts are benign PyTorch calls, Markdown formatting, dataset cache paths, dictionary keys, imports, or official URLs.

Files scanned

3,566

Lines analyzed

Review items

False positives ignored

Capability review items (1)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Medium

scripts/create_gnn_template.py:467

Python file write/append

with open(output_file, 'w') as f:

The CLI-controlled output path is created and opened with write mode without workspace or existence checks, allowing overwrite of any writable file.

Risk Factors

📁 Filesystem access (50)

⚡ Contains scripts (8)

scripts/benchmark_model.py:97 scripts/benchmark_model.py:129 scripts/create_gnn_template.py:72 scripts/create_gnn_template.py:168 scripts/create_gnn_template.py:265 scripts/create_gnn_template.py:392 SKILL.md:350 SKILL.md:630

⚙️ External commands (50)

🌐 Network access (5)

SKILL.md:33 SKILL.md:667 SKILL.md:668 SKILL.md:669 SKILL.md:670

Audited by: codex View Audit History →

Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report

Copy report link

https://skillstore.io/skills/davila7-torch-geometric/audits/11?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/davila7-torch-geometric/security.svg)](https://skillstore.io/skills/davila7-torch-geometric?utm_source=security_passport_badge)

HTML badge

<a href="https://skillstore.io/skills/davila7-torch-geometric?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/davila7-torch-geometric/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Embed card

<iframe src="https://skillstore.io/embed/skills/davila7-torch-geometric.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Academic citations (APA · BibTeX · CFF)

APA citation

davila7. (2026). torch-geometric security audit report (audit version 11) [Author version unspecified]. Skillstore. https://skillstore.io/skills/davila7-torch-geometric/audits/11

BibTeX citation

@techreport{davila7-davila7-torch-geometric-2026,
  author = {davila7},
  title = {torch-geometric security audit report (audit version 11)},
  institution = {Skillstore},
  year = {2026},
  number = {11},
  url = {https://skillstore.io/skills/davila7-torch-geometric/audits/11},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "torch-geometric security audit report (audit version 11)"
version: "unspecified"
type: report
authors:
  - name: "davila7"
date-released: "2026-07-23"
url: "https://skillstore.io/skills/davila7-torch-geometric/audits/11"
identifiers:
  - type: other
    value: "skillstore:davila7-torch-geometric:audit:11"
    description: "Skillstore immutable audit report identifier"

Compare variants

2 installable variants

Each author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.

Why this variant is first

Highest Skillstore Score

Author / source Skillstore Score Evidence Confidence Skillstore usage Updated

davila7 Recommended Current

davila7-torch-geometric

Skillstore Score 80

Evidence Confidence High

Skillstore usage 10

Updated

2026-08-21

K-Dense-AI

k-dense-ai-torch-geometric

Skillstore Score 72

Evidence Confidence High

Skillstore usage 11

Updated

2026-08-21

Skillstore Score

Why this score Evidence Confidence: High

Architecture

Maintainability

Content

Community

Spec Compliance

What You Can Build

Learn graph modeling fundamentals

Create Data objects, understand edge indices, and train a first node classifier with clear tensor and evaluation guidance.

Prototype scalable GNN workflows

Compare architectures, apply transforms, use neighbor sampling, and benchmark models before integrating them into a larger project.

Model scientific graph data

Select datasets, layers, pooling, and heterogeneous graph patterns for molecular, biological, spatial, or relational research.

Try These Prompts

Create a basic PyG graph

Represent my node features and edge list as a PyTorch Geometric Data object. Explain tensor shapes and validation checks.

Build a node classifier

Design a GCN node classification workflow for [dataset]. Include splits, preprocessing, training, validation, testing, and metrics.

Scale graph training

Adapt my [model] for a graph with [node count] nodes. Choose neighbor sampling, batch settings, device handling, and memory controls.

Design a heterogeneous pipeline

Design a heterogeneous GNN for [node types], [edge types], and [target]. Compare conversion and HeteroConv, then define evaluation and explainability.

Best Practices

Validate feature shapes, edge indices, labels, masks, and graph direction before training.
Use fixed splits and seeds, then report repeated runs with hardware and package versions.
Start with a simple baseline before adding attention, deeper layers, sampling, or heterogeneous relations.

Avoid

Do not evaluate on the test split while selecting models or hyperparameters.
Do not load a full large graph onto one device when neighbor sampling is required.
Do not compare architectures with different data splits, preprocessing, or training budgets.

Frequently Asked Questions

Does this skill install PyTorch Geometric?

No. It provides installation commands, but users must run them and select versions compatible with their PyTorch and accelerator setup.

Which tasks are covered?

The guidance covers node classification, graph classification, link prediction, heterogeneous graphs, molecular modeling, geometric data, and large-graph sampling.

Can it generate complete training code?

Yes, for selected GCN, GAT, GraphSAGE, and GIN examples. Generated templates still require review and adaptation.

Does it support custom datasets?

Yes. It explains Data objects, in-memory and disk-backed datasets, CSV loading, transforms, and split construction.

Can it benchmark models?

Yes. The bundled benchmark compares supported architectures on common node or graph datasets and summarizes accuracy and training time.

Can it visualize graphs?

Yes. The visualization utility can render graph structure, labels, attributes, degree distributions, and summary statistics for supported datasets.