mood-board-builder

I took a look at your mood-board-builder skill and noticed the PDA score pulled things down significantly—you've got solid foundational logic, but the way you're structuring the examples and guidance could be way more token-efficient for users actually building with this.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 58/100, which lands you in F territory. This is based on Anthropic's best practices for progressive disclosure architecture and utility. Your strongest pillar is Spec Compliance (11/15)—the YAML frontmatter is solid—but Utility (10/20) and PDA (15/30) are dragging the overall score down. The skill addresses a real problem, but it reads more like a concept than something Claude can actually execute.

What's Working Well

• Valid structure: Your frontmatter is clean and follows conventions correctly
• Clear workflow intent: The 5-step framework (Inspiration → Direction → Refinement → Feedback → Finalization) makes conceptual sense
• Real use case: Mood boards are genuinely useful for creative alignment in teams
• Navigation: Sections are well-labeled and easy to scan

The Big One: No Actual Templates or Examples

This is killing your Utility score. You mention "Figma/Canva board layout, PDF deck template" but provide zero actual artifacts. When Claude uses this skill, it has no concrete reference for what a "mood board" output should actually look like.

Fix: Create a references/ directory with:

• references/sample-mood-board.md - A fully formatted markdown example showing vision statement, 3 color palettes (with hex codes), typography selections, and annotated reference images
• references/feedback-template.md - Actual stakeholder feedback form users can paste
• references/figma-template-link.txt - Direct link to a Figma community file

This moves you from describing mood boards to showing Claude exactly what to generate. Estimate: +5 points on Utility alone.

Other Things Worth Fixing

1. Missing trigger phrases in description (Spec Compliance issue): Your frontmatter description says "compile visual inspiration" but doesn't include keywords like "mood board", "visual direction", or "creative concepting". Add these to the description field so discoverability works. +2 points

2. Framework steps need actionable outputs: "Inspiration Sourcing – collect references" is vague. Make it concrete: "Generate 15-20 specific reference suggestions with source URLs and why each matters for the brand." +3 points

3. No feedback/iteration loop: You mention stakeholder feedback forms exist but never explain how Claude validates the board quality or iterates. Add a validation step: "Ask: Does this align with brand guidelines? Are the 3 directions visually distinct? Check color contrast ratios for accessibility." +2 points

Quick Wins

• Add actual template files to references/ directory (+5 pts)
• Include specific trigger keywords in SKILL.md metadata (+2 pts)
• Replace conceptual steps with concrete outputs Claude should produce (+3 pts)
• Add a validation/iteration section (+2 pts)

These four changes should bump you to mid-70s territory and make the skill actually usable.

Checkout your skill here: SkillzWave.ai | SpillWave We have an agentic skill installer that install skills in 14+ coding agent platforms. Check out this guide on how to improve your agentic skills.