senior-architect

I noticed your skill focuses on architectural patterns, but the grading came back at 28/100—let me walk through what's holding it back and where we can rebuild this to actually help developers make better design decisions.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 28/100, which puts this in F territory. This is based on Anthropic's best practices for skill design. Your strongest area is Spec Compliance (12/15)—the YAML frontmatter is clean—but Utility is completely bottlenecked at 0/20, dragging everything down. The core issue: the skill promises tools and workflows that don't actually exist.

What's Working Well

• Valid YAML structure - Your frontmatter hits all the required fields with correct hyphen-case naming
• Clear metadata description - You nailed the tech stack details and trigger phrases in the description
• Organized sections - Headers for Features, Workflows, and References show intentional structure

The Big One: Missing Core Files and Scripts

This is the wall stopping everything. Your SKILL.md references three documentation files (architecture_patterns.md, system_design_workflows.md, tech_decision_guide.md) and three Python scripts (architecture_diagram_generator.py, project_architect.py, dependency_analyzer.py) that don't exist.

When a developer tries to use this skill, they hit dead links and broken commands. That's why Utility scored 0/20—there's no actual problem-solving power here.

The fix: Either create these files with real content, or consolidate everything into SKILL.md itself. You don't need external scripts—replace them with concrete Claude workflows. For example, instead of referencing a missing diagram generator, show an actual prompt pattern developers can use with Claude to generate architecture diagrams. This alone could jump you 8-10 points.

Other Things Worth Fixing

1. Strip the marketing language - "comprehensive toolkit," "expert-level automation," "deep analysis"—these add fluff. Replace with concrete capabilities: "Analyze component coupling and suggest decoupling strategies" beats "deep analysis" every time. (+3 points)

2. Build real workflows with decision trees - Your "Development Workflow" section just says "Follow the patterns"—replace with numbered steps that actually guide someone through design decisions (choose database strategy, define API contracts, etc.). (+5 points)

3. Replace vague feature lists - "Performance metrics," "Automated fixes," "Recommendations" tell me nothing. Make them concrete or cut them. (+4 points)

4. Add a table of contents - At 210 lines, this needs a TOC at the top so developers can skip to what matters to them. (+2 points)

Quick Wins

• Highest impact: Create or inline the three missing reference files with real architecture patterns and system design workflows
• Next: Remove marketing fluff, replace with concrete problem statements
• Then: Build actual decision trees in your workflows section
• Finally: Add table of contents and consistent terminology throughout

This rewrite could realistically get you to 60-70/100 if you focus on making it real and actionable instead of aspirational.

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.