performance-tracking

I noticed you're tackling performance tracking—it's one of those problems that seems simple until you start handling real-world edge cases. Your implementation scores at 70/100 (Grade C), so there's solid foundation here, but I spotted some opportunities where the progressive disclosure could be tightened and a few rough edges in how developers interact with the core APIs.

Links:

• GitHub
• SkillzWave Marketplace

The TL;DR

You're sitting at 70/100, which is solid C-territory (70-79 is "adequate, with gaps"). This is based on Anthropic's skill best practices framework. Your strongest area is Spec Compliance (12/15) — the frontmatter and naming conventions are locked in. Weakest is Progressive Disclosure (18/30) — you've put everything into SKILL.md when reference files would chunk it down and save tokens for users.

What's Working Well

• Frontmatter and naming — Valid YAML with all required fields, proper hyphen-case format. No friction there.
• Clear logical flow — Your framework steps (Metric Hierarchy → Dashboard → Monitoring → Optimization) follow a sensible progression that matches how people actually think about measurement.
• Consistent terminology — You're not switching between "KPI" and "metric" randomly, which saves cognitive load.
• grep-friendly structure — The skill has a clean, scannable format (+1 bonus point).

The Big One: Missing Reference Architecture

Here's what's holding you back the most: everything is stuffed into SKILL.md. You mention templates (KPI tree worksheet, dashboard spec, optimization log) but they're just labels with no actual content. This kills your PDA score because users pay token cost for all that text, but get no working artifact.

The fix: Create a references/ direc...