I took a look at your FDA consultant skill and noticed the grading framework focuses heavily on regulatory compliance patterns—but I'm curious whether the skill documentation assumes users already know which FDA workflows they're targeting, or if it guides them through that discovery first.
Links:
The TL;DR
You're at 58/100, which lands in F territory. This is based on Anthropic's best practices for skill architecture. Your strongest area is Spec Compliance (12/15)—the frontmatter and naming conventions are solid. The real drag is Writing Style (4/10) and Progressive Disclosure Architecture (13/30)—lots of marketing fluff and missing reference files are eating your score.
What's Working Well
- Valid YAML structure — Your frontmatter is clean and follows conventions properly
- Good trigger identification — You nailed the discovery terms: "FDA submission", "QSR compliance", "HIPAA" are all solid FDA-specific triggers
- Structured sections — The skill has clear headers and navigation signals that help users find what they need
- Addresses real problems — FDA compliance is genuinely useful, and your workflows target actual regulatory pain points
The Big One: Missing Reference Architecture
Here's the problem: Your SKILL.md references 9+ files that don't exist (fda-capa-requirements.md, device-cybersecurity-guidance.md, software-hipaa-compliance.md, etc.), but only a placeholder file actually exists in the references directory. This is a -7 point hit because it breaks Progressive Disclosure Architecture—users click expecting detailed guidance and find nothing.
Fix: Either (1) create all referenced files with actual content—QSR requirements, HIPAA frameworks, submission guides—or (2) consolidate everything into SKILL.md with inline sections. The skill is 246 lines; you have room. This alone gets you +7 points.
Other Things Worth Fixing
-
Trim the marketing language — Your intro reads like a LinkedIn profile ("Expert-level", "comprehensive knowledge", "Senior FDA Consultant"). Replace with instructional voice: "Provide FDA regulatory consulting for medical device companies" instead. That's +3 points.
-
Cut verbose elaboration — Sentences like "Provide expert guidance on optimal FDA regulatory pathways ensuring efficient market access and regulatory compliance" should be "Guide FDA pathway selection for efficient market access." Your skill is losing token economy points because every line needs to earn its tokens. +4 points if you tighten this throughout.
-
Add validation checkpoints — Your workflows list steps but no verification. Add: "2. VERIFY: Confirm predicate device substantial equivalence against FDA checklist" after each major step. This adds feedback loops Utility needs. +3 points.
-
Standardize workflow format — Some sections use numbered steps, others use nested bullets. Pick one and stick with it. Also flag decision points consistently ("Decision Point: Evaluate..."). +2 points.
Quick Wins
- Create the missing reference files (or consolidate into SKILL.md) → +7 points
- Replace marketing language with instructional tone → +3 points
- Tighten verbose sentences for token economy → +4 points
- Add verification steps to workflows → +3 points
- Standardize workflow numbering → +2 points
That's +19 points within reach—gets you to 77/100 with relatively focused changes.
Checkout your skill here: [SkillzWave.ai](https://skillzwave.ai) | [SpillWave](https://spillwave.com) 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.
I took a look at your FDA consultant skill and noticed the grading framework focuses heavily on regulatory compliance patterns—but I'm curious whether the skill documentation assumes users already know which FDA workflows they're targeting, or if it guides them through that discovery first.
Links:
The TL;DR
You're at 58/100, which lands in F territory. This is based on Anthropic's best practices for skill architecture. Your strongest area is Spec Compliance (12/15)—the frontmatter and naming conventions are solid. The real drag is Writing Style (4/10) and Progressive Disclosure Architecture (13/30)—lots of marketing fluff and missing reference files are eating your score.
What's Working Well
The Big One: Missing Reference Architecture
Here's the problem: Your SKILL.md references 9+ files that don't exist (
fda-capa-requirements.md,device-cybersecurity-guidance.md,software-hipaa-compliance.md, etc.), but only a placeholder file actually exists in the references directory. This is a -7 point hit because it breaks Progressive Disclosure Architecture—users click expecting detailed guidance and find nothing.Fix: Either (1) create all referenced files with actual content—QSR requirements, HIPAA frameworks, submission guides—or (2) consolidate everything into SKILL.md with inline sections. The skill is 246 lines; you have room. This alone gets you +7 points.
Other Things Worth Fixing
Trim the marketing language — Your intro reads like a LinkedIn profile ("Expert-level", "comprehensive knowledge", "Senior FDA Consultant"). Replace with instructional voice: "Provide FDA regulatory consulting for medical device companies" instead. That's +3 points.
Cut verbose elaboration — Sentences like "Provide expert guidance on optimal FDA regulatory pathways ensuring efficient market access and regulatory compliance" should be "Guide FDA pathway selection for efficient market access." Your skill is losing token economy points because every line needs to earn its tokens. +4 points if you tighten this throughout.
Add validation checkpoints — Your workflows list steps but no verification. Add: "2. VERIFY: Confirm predicate device substantial equivalence against FDA checklist" after each major step. This adds feedback loops Utility needs. +3 points.
Standardize workflow format — Some sections use numbered steps, others use nested bullets. Pick one and stick with it. Also flag decision points consistently ("Decision Point: Evaluate..."). +2 points.
Quick Wins
That's +19 points within reach—gets you to 77/100 with relatively focused changes.
Checkout your skill here: [SkillzWave.ai](https://skillzwave.ai) | [SpillWave](https://spillwave.com) 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.