Every skill is scored on 6 dimensions
SupaScore is our transparent quality scoring system. Each skill earns its tier through rigorous, weighted evaluation.
Average Score
84.1
Total Scored
1000
—
139
861
—
The 6 dimensions
Research Quality
15%Depth and accuracy of the research backing this skill. Cited sources, peer-reviewed references, authoritative standards.
Prompt Engineering
25%Quality of the system prompt itself. Clear instructions, structured output, guard rails, parameter handling.
Practical Utility
15%Real-world applicability. Does the skill produce actionable, implementable output?
Completeness
10%Coverage of the topic. Edge cases, trade-offs, alternatives, and limitations addressed.
User Satisfaction
20%How useful users find the skill in practice. Based on ratings and feedback.
Decision Usefulness
15%Does the output help make better decisions? Not just information — actionable intelligence.
The formula
SupaScore = (RQ × 0.15) + (PE × 0.25) + (PU × 0.15) + (CO × 0.10) + (US × 0.20) + (DU × 0.15)Each dimension scored 0–10. Weighted sum produces a 0–100 score. Tier is assigned based on thresholds.
Tier thresholds
95 – 100
Exceptional. Near-perfect across all dimensions.
85 – 94
Excellent. Strong evidence and prompt quality.
70 – 84
Solid. Well-researched and practical.
60 – 69
Good foundation. Room for improvement.
< 60
Below publication threshold. Not visible in catalog.
Evidence policy
Skills that handle sensitive domains (finance, legal, medical) have an evidence policy flag. When enabled, the system prompt explicitly requires Claude to cite sources and distinguish between evidence-based claims and opinions.
Risk domains trigger warning badges on skill detail pages. High-risk skills may require human sign-off before acting on outputs.