Content
77%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A strong skill body with excellent workflow structure and best-practice progressive disclosure via a load-conditions reference table. Weaknesses are concentrated in token efficiency: a name-dump "Knowledge Reference" section, scattered pinned version numbers, and one or two code lines that would not execute as written.
Suggestions
Delete the "Knowledge Reference" name-list section and the maintainer/footer credit lines — they restate technologies Claude already knows and add no actionable guidance.
Consolidate pinned versions (Storefront API 2024-10, Shopify CLI 3.x, App Bridge 4.0, Hydrogen 2024, Liquid 2.0) into a single "Versions / compatibility" note (or move them into the reference files) so the main body stays version-agnostic and does not silently go stale.
Fix the non-executable code lines: replace `{{ product.description | metafield_tag }}` with `{{ product.metafields.custom.care_instructions | metafield_tag }}`, and replace `shopify app create node` with the CLI 3.x scaffold command (`npm init @shopify/create-app@latest`) or drop it in favor of the existing reference.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient — code examples earn their tokens — but the "Knowledge Reference" section is a bare name-dump ("Shopify CLI 3.x, Liquid 2.0, ... Hydrogen 2024, Remix, Oxygen, Polaris, App Bridge 4.0...") of technologies Claude already knows, and pinned time-sensitive versions ("Storefront API 2024-10 or newer", "Hydrogen 2024", "App Bridge 4.0", "Shopify CLI 3.x") are scattered in constraints rather than isolated in a versioning/deprecated section. The footer credit links also add no operational value. Fits anchor 3 ("mostly efficient but includes some unnecessary explanation or could be tightened"). Not 4 because the version pins and the name-list padding are more than trimmable one-liners; not 2 because the bulk (code, commands, reference table) is dense and useful. | 3 / 5 |
Actionability | Five copy-paste-ready blocks cover the common cases: Liquid product template with metafields, OS 2.0 collection filters, a Storefront GraphQL query, CLI command list, and a TypeScript authenticated-admin loader. However, minor gaps exist: `{{ product.description | metafield_tag }}` applies a metafield filter to a plain string, and "shopify app create node" is not a CLI 3.x command (scaffolding is via npm init), so a couple of lines would not run as-is. Fits anchor 4 ("mostly executable guidance; concrete code or commands with minor gaps"). Not 5 because of those broken/outdated lines; not 3 because everything is real code and commands, not pseudocode. | 4 / 5 |
Workflow Clarity | The five-step Core Workflow has an explicit validation step with a feedback loop: "Run `shopify theme check` for Liquid linting; if errors are found, fix them and re-run before proceeding... If validation fails at any step, resolve all reported issues before moving to deployment". Matches anchor 5 (clear sequence, explicit validation, error-recovery loop). Not 4 because validation is not merely present but includes the fix-and-retry loop and a gate before deployment. | 5 / 5 |
Progressive Disclosure | The Reference Guide table maps each of the five topics to an existing one-level-deep file in references/ with an explicit "Load When" condition (e.g. "`references/liquid-templating.md` | Theme development, template customization"); all five files exist and are substantive (690–936 lines). SKILL.md stays an overview with core examples while bulk detail lives in references. Matches anchor 5. Not 4 because navigation is explicit, references are one level deep, and the split is clean with no orphaned or buried paths. | 5 / 5 |
Total | 17 / 20 Passed |