Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create the deliverable (code, documentation, tests, content) following the user's standards and best practices. Use after validation passes to actually build the work product.
.claude/skills/aiskillstore-generate-output/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 148% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 215% | 0% |
Transforms validated requirements into actual deliverables (code, documentation, tests, content). This skill follows the user's principles and common patterns defined in their standards.
Use StandardsRepository to access standards:
javascriptconst standards = standardsRepository.getStandards(context.projectType) if (standards) { // Use their principles and patterns const principles = standards.principles const patterns = standards.commonPatterns // Generate output following their standards } else { // Generate following best practices }
See .claude/lib/standards-repository.md for interface details.
Deliver the work with:
# [Title of Deliverable]
## Summary
[Brief description of what was created]
## Principles Applied
- [First principle from their standards]
- [Second principle]
- [Third principle]
## Common Patterns Included
- [Pattern 1: brief explanation]
- [Pattern 2: brief explanation]
## The Deliverable
[Complete code, documentation, tests, or content]
## Next Steps
[What they should do next - formatting, testing, review]✓ Deliverable is complete and ready to use ✓ Follows user's principles and patterns ✓ Appropriate to project type ✓ Professional quality ✓ Includes necessary supporting elements (comments, examples, structure)
Project Type: React Component User Requirements: "Searchable dropdown component with keyboard nav" Their Standards Include:
Generated Output Includes:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 17,677 | 16,034 | -9% | 1 | 1 | 0% | 3,893 | 4,128 | +6% | 0 | 0 | — |
case-02 | fail→pass | 20,416 | 20,062 | -2% | 1 | 1 | 0% | 4,446 | 5,282 | +19% | 0 | 0 | — |
case-03 | pass→pass | 8,047 | 14,388 | +79% | 1 | 1 | 0% | 1,330 | 3,305 | +148% | 0 | 0 | — |
case-04 | fail→fail | 5,005 | 14,416 | +188% | 1 | 1 | 0% | 1,001 | 3,641 | +264% | 0 | 0 | — |
case-05 | pass→pass | 3,660 | 6,519 | +78% | 1 | 1 | 0% | 644 | 2,030 | +215% | 0 | 0 | — |
case-06 | pass→pass | 15,836 | 22,102 | +40% | 1 | 1 | 0% | 3,468 | 5,711 | +65% | 0 | 0 | — |
case-07 | pass→pass | 27,941 | 25,557 | -9% | 1 | 1 | 0% | 6,189 | 7,124 | +15% | 0 | 0 | — |
case-08 | pass→pass | 13,473 | 15,186 | +13% | 1 | 1 | 0% | 2,889 | 4,214 | +46% | 0 | 0 | — |
case-09 | fail→pass | 19,655 | 19,471 | -1% | 1 | 1 | 0% | 3,650 | 4,403 | +21% | 0 | 0 | — |
case-10 | pass→pass | 19,402 | 19,584 | +1% | 1 | 1 | 0% | 3,825 | 5,203 | +36% | 0 | 0 | — |
case-11 | pass→pass | 17,105 | 19,483 | +14% | 1 | 1 | 0% | 3,544 | 5,210 | +47% | 0 | 0 | — |
case-12 | pass→pass | 22,620 | 20,207 | -11% | 1 | 1 | 0% | 5,240 | 5,409 | +3% | 0 | 0 | — |
case-13 | pass→pass | 22,591 | 23,939 | +6% | 1 | 1 | 0% | 3,759 | 4,993 | +33% | 0 | 0 | — |
case-14 | pass→pass | 17,802 | 18,864 | +6% | 1 | 1 | 0% | 2,956 | 4,008 | +36% | 0 | 0 | — |
case-15 | fail→fail | 11,994 | 16,456 | +37% | 1 | 1 | 0% | 2,251 | 4,524 | +101% | 0 | 0 | — |
case-16 | pass→pass | 12,013 | 14,110 | +17% | 1 | 1 | 0% | 2,474 | 3,806 | +54% | 0 | 0 | — |
case-17 | pass→pass | 18,622 | 16,847 | -10% | 1 | 1 | 0% | 3,789 | 4,432 | +17% | 0 | 0 | — |
case-18 | pass→pass | 19,849 | 25,752 | +30% | 1 | 1 | 0% | 4,294 | 6,556 | +53% | 0 | 0 | — |
case-19 | pass→pass | 15,458 | 15,103 | -2% | 1 | 1 | 0% | 3,347 | 4,200 | +25% | 0 | 0 | — |
case-20 | pass→pass | 12,323 | 15,846 | +29% | 1 | 1 | 0% | 2,436 | 4,241 | +74% | 0 | 0 | — |
case-21 | pass→pass | 18,009 | 21,972 | +22% | 1 | 1 | 0% | 3,747 | 5,970 | +59% | 0 | 0 | — |
case-22 | pass→pass | 11,857 | 16,432 | +39% | 1 | 1 | 0% | 2,768 | 4,743 | +71% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.