torch-geometric
Graph Neural Networks (PyG). Node/graph classification, link prediction, GCN, GAT, GraphSAGE, heterogeneous graphs, molecular property prediction, for geometric deep learning.
Install with Tessl CLI
npx tessl i github:K-Dense-AI/claude-scientific-skills --skill torch-geometricOverall
score
79%
Does it follow best practices?
If you maintain this skill, you can automatically optimize it using the tessl CLI to improve its score:
npx tessl skill review --optimize ./path/to/skillValidation for skill structure
Discovery
83%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong technical description with excellent specificity and trigger term coverage for the graph neural networks domain. The main weakness is the absence of an explicit 'Use when...' clause, which would help Claude know exactly when to select this skill. The description correctly uses third person voice and avoids vague language.
Suggestions
Add an explicit 'Use when...' clause, e.g., 'Use when the user asks about graph neural networks, PyTorch Geometric, node embeddings, or graph-based machine learning tasks.'
Consider adding common user phrasings like 'graph ML', 'message passing', or 'graph embeddings' to capture more natural language variations.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions and architectures: 'Node/graph classification, link prediction, GCN, GAT, GraphSAGE, heterogeneous graphs, molecular property prediction' - these are specific, actionable capabilities. | 3 / 3 |
Completeness | Clearly answers 'what does this do' with specific capabilities, but lacks an explicit 'Use when...' clause or equivalent trigger guidance. The 'when' is only implied through the domain terms listed. | 2 / 3 |
Trigger Term Quality | Excellent coverage of natural terms users would say: 'Graph Neural Networks', 'PyG', 'GCN', 'GAT', 'GraphSAGE', 'node classification', 'link prediction', 'molecular property prediction', 'geometric deep learning' - these are exactly what practitioners would search for. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche - Graph Neural Networks with PyTorch Geometric is a very specific domain. The combination of PyG, specific architectures (GCN, GAT, GraphSAGE), and tasks (node classification, link prediction) makes it unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
73%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a comprehensive and highly actionable PyTorch Geometric skill with excellent code examples covering the full spectrum from basic usage to advanced heterogeneous graphs. The main weaknesses are some verbosity in explanatory sections that Claude doesn't need, missing validation/verification steps in workflows, and an unnecessary promotional section at the end that wastes tokens.
Suggestions
Remove the promotional K-Dense Web section at the end - it adds no value to the skill and wastes context window tokens
Add validation checkpoints to training workflows (e.g., 'Verify data.x shape matches expected dimensions before training', 'Check for NaN losses during training')
Trim explanatory text that describes concepts Claude already knows (e.g., the detailed explanation of mini-batch processing mechanics, what GNNs are)
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is comprehensive but includes some unnecessary explanations (e.g., explaining what GNNs do, describing mini-batch processing concepts Claude already knows). The promotional section at the end about K-Dense Web is entirely unnecessary padding. | 2 / 3 |
Actionability | Excellent executable code examples throughout - from basic graph creation to full training workflows for node classification, graph classification, and large-scale graphs. All code is copy-paste ready with proper imports. | 3 / 3 |
Workflow Clarity | Training workflows are presented as clear sequences, but lack explicit validation checkpoints. No verification steps for model saving/loading, no error handling patterns, and no feedback loops for debugging common issues like dimension mismatches. | 2 / 3 |
Progressive Disclosure | Well-structured with clear sections progressing from quick start to advanced features. References to external files (references/layers_reference.md, scripts/) are clearly signaled and one level deep. Content is appropriately split between overview and detailed references. | 3 / 3 |
Total | 10 / 12 Passed |
Validation
81%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 13 / 16 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
skill_md_line_count | SKILL.md is long (676 lines); consider splitting into references/ and linking | Warning |
description_trigger_hint | Description may be missing an explicit 'when to use' trigger hint (e.g., 'Use when...') | Warning |
metadata_version | 'metadata.version' is missing | Warning |
Total | 13 / 16 Passed | |
- Reviewed
Table of Contents
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.