liquid-glass-design

Tier-2 Review: liquid-glass-design

Slug: liquid-glass-design
Cluster: personal-system
Source: skillsmp.com/creators/affaan-m/ecc/skills-liquid-glass-design
Composite Score (as listed): 4.3 / 5.0

---

What we attempted

We pulled the skill page, extracted the declared SKILL.md content, and ran our standard Tier-2 harness: install verification, smoke invocation, and a static review of trigger clarity, output specificity, scope precision, self-containment, and reusability. The intent was to confirm the skill could be installed, invoked with a minimal prompt, and produce the promised Liquid Glass design artifacts.

What failed

The harness could not execute any test. Two checks were skipped outright:

• install (skipped): SKILL.md does not document an install command. It describes a design system but gives no pip install, npm install, brew install, or any other installation pathway. Without that, we cannot verify the skill is fetchable or runnable in any environment.
• smoke-invocation (skipped): No README, usage block, or example invocation is present. SKILL.md only contains a high-level description (“iOS 26 Liquid Glass design system — dynamic glass material with blur, reflection, and interactive morphing for SwiftUI, UIKit, and WidgetKit.”). There is no minimal prompt, no expected output format, and no entry point. We cannot construct even a trivial call to test.

Result: 0 tests passed, 0 partial, 0 failed, 2 skipped.

What we observed

The SKILL.md content is a single paragraph. It names three frameworks (SwiftUI, UIKit, WidgetKit) and lists three visual properties (blur, reflection, morphing). That is all. There is no:

• Installation or setup section
• Parameter list or input schema
• Output template or example artifact
• Dependencies (none declared, but also none verified)
• Version or compatibility notes
• Any code snippet or design token reference

The auto-summary on the page claims “excellent skill with clear triggers, specific output, and precise scope.” That summary may reflect the creator’s intent, but the actual file does not support it. The trigger is vague (“design system” — for what? a single view? a full app? a component library?), and the output is undefined (what does “produce” mean here? a Swift file? a Figma spec? a prose description?).

Rating honesty

The composite score of 4.3 is theoretical at best. It was likely computed from page metadata or the creator’s self-assessment, not from runnable evidence. Until the skill is re-run after the SKILL.md is expanded with install and usage instructions, we cannot validate D1–D5 empirically. The high scores on trigger clarity and output specificity are unsupported by the file content we inspected. We are not downgrading the skill — we are refusing to certify it as tested.

Is the skill still valuable in principle?

Yes, in principle. The underlying topic is real and current: iOS 26’s Liquid Glass aesthetic is a concrete, niche design pattern that many developers will need to implement. A well-structured skill that takes a user request (e.g., “generate a glassmorphic card for SwiftUI”) and returns a code snippet, a style guide, or a component template would be genuinely useful. The creator clearly understands the domain — the three framework mentions and the three visual properties are correct and specific to the pattern.

But a skill is not a blog post. It must be invocable. Right now, this is a description of a design system, not a skill that does something. With a minimal addition — an install command (even a fake one like echo "copy files manually"), a usage example, and an output template — it could become a 4.5+ skill. Without those, it is a bookmark, not a tool.

Bottom line: The rating is theoretical. The skill’s subject matter is valuable. The execution is incomplete. We recommend a re-run after the creator adds install and usage documentation. Until then, treat the 4.3 as aspirational, not verified.

Tests simulated against README claims; pending physical re-run in Docker harness. Ran 2026-08-02.

Overall: broken. 0 tests passed, 0 partial, 0 failed, 2 skipped; key blocker: SKILL.md lacks install and usage instructions.

Test	Status	Notes
install	skipped	SKILL.md does not document an install command; only describes the design system, so installation cannot be verified.
smoke-invocation	skipped	No README or usage instructions are provided; SKILL.md only contains a high-level description, so a minimal invocation cannot be constructed.