torch-geometric

Graph Neural Networks (PyG). Node/graph classification, link prediction, GCN, GAT, GraphSAGE, heterogeneous graphs, molecular property prediction, for geometric deep learning.

Quality

73%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./bundled/skills/torch-geometric/SKILL.md

Quality

Content

65%Weight 40%Scale 1-3

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

The body is highly actionable with extensive executable PyG code and a reasonable one-level reference structure, but it is verbose in places, lacks validation feedback loops in its batch training workflows, and signals two reference files that are missing from the bundle.

Suggestions

Remove or create the missing referenced files: "references/layer_capabilities.md" (line ~652) and "references/api_patterns.md" (line ~663) are cited but absent from references/, leaving broken links.

Trim conceptual padding that Claude already knows (Overview definition of PyG, the four-step message-passing paradigm) and consolidate detail already covered in the bundled reference files.

Add a validation/checkpoint step to at least the batch training and neighbor-sampling workflows (e.g., assert expected tensor shapes or check loss/accuracy before continuing) to satisfy the feedback-loop expectation for batch operations.

Dimension	Reasoning	Score
Conciseness	The body is mostly actionable PyG-specific material, but includes conceptual padding Claude already knows (the Overview's "library built on PyTorch for developing and training Graph Neural Networks", the message-passing paradigm steps) and overlaps content already in the bundled reference files; it could be tightened.	2 / 3
Actionability	Abundant concrete, copy-paste-ready Python covering graph creation, dataset loading, GCN/GAT/GraphSAGE layers, custom message passing, custom datasets, and full training workflows, matching the fully-executable anchor.	3 / 3
Workflow Clarity	Training workflows (node classification, graph classification, neighbor sampling) are clearly sequenced, but none include validate→fix→retry feedback loops for the batch DataLoader/NeighborLoader operations, which the rubric notes cap at 2.	2 / 3
Progressive Disclosure	Structure is one level deep with inline-signaled references, but two referenced files do not exist in the bundle ("references/layer_capabilities.md" and "references/api_patterns.md"), so navigation breaks; this prevents the "easy navigation" bar of a 3.	2 / 3
	Total	9 / 12 Passed

Description

82%Weight 40%Scale 1-3

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

The description is specific and trigger-rich for a well-defined GNN niche, but it omits an explicit "Use when…" clause, leaving the activation guidance implicit rather than stated.

Suggestions

Add an explicit trigger clause such as "Use when working with graph neural networks, torch_geometric/PyG, or graph-based deep learning tasks." to answer both what and when.

Tighten the trailing fragment "for geometric deep learning" into a complete clause so the sentence reads cleanly rather than as a dangling purpose phrase.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete tasks — "Node/graph classification", "link prediction", "molecular property prediction" — alongside specific architectures (GCN, GAT, GraphSAGE), matching the multiple-specific-actions anchor.	3 / 3
Completeness	It states what the skill does but has no explicit "Use when…" trigger clause; the "when" is only implied by the trailing "for geometric deep learning", so per the judging guideline completeness is capped at 2.	2 / 3
Trigger Term Quality	Strong coverage of terms a GNN user would naturally say, including the common abbreviation "PyG" and concrete model names "GCN, GAT, GraphSAGE", plus "heterogeneous graphs" and "geometric deep learning".	3 / 3
Distinctiveness Conflict Risk	The Graph Neural Network / PyG niche is clearly distinct with specialized triggers unlikely to overlap with other skills.	3 / 3
	Total	11 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (681 lines); consider splitting into references/ and linking	Warning
referenced_paths_exist	Referenced path issues: 2 missing	Warning

	Total	14 / 16 Passed

Repository: foryourhealth111-pixel/Vibe-Skills
Path: bundled/skills/torch-geometric/SKILL.md
Commit: f627ab5

Reviewed: 11 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.