Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate canonical FrontendIntegration YAML from a simplified single-menu authoring model for frontend-forge.
.claude/skills/kubesphere-frontend-integration-yaml/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 14% | 0% |
Generate submit-ready FrontendIntegration YAML from a simplified authoring model with one menu and one or more pages.
Ask the user instead of guessing when any of these are missing:
metadata.namemenu.displayNamemenu.placements[]pages[]crdTable: group, version, scope, names.pluraliframe: key and iframe.srcUse the simplified single-menu model as input to the generator:
metadataspec for optional passthrough fieldsmenupagesDo not author canonical spec.menus[] by hand.
scripts/generate_frontend_integration.py.kind: FrontendIntegration.menu.placements[] into spec.menus[].children from pages[].workspace, force all crdTable.scope values to Namespaced.menu.icon to GridDuotone when it is omitted.yaml code block.FrontendIntegration YAML has been generated. Reply 'apply' to deploy it directly.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-20 | fail→fail | 11,895 | 6,468 | -46% | 1 | 1 | 0% | 2,224 | 1,658 | -25% | 0 | 0 | — |
case-01 | fail→pass | 7,753 | 5,020 | -35% | 1 | 1 | 0% | 1,504 | 1,653 | +10% | 0 | 0 | — |
case-02 | pass→pass | 7,177 | 4,255 | -41% | 1 | 1 | 0% | 1,222 | 1,483 | +21% | 0 | 0 | — |
case-03 | fail→pass | 8,901 | 5,895 | -34% | 1 | 1 | 0% | 1,604 | 1,762 | +10% | 0 | 0 | — |
case-04 | fail→pass | 6,270 | 7,172 | +14% | 1 | 1 | 0% | 1,115 | 1,952 | +75% | 0 | 0 | — |
case-05 | fail→pass | 8,092 | 2,957 | -63% | 1 | 1 | 0% | 1,374 | 1,042 | -24% | 0 | 0 | — |
case-06 | fail→fail | 9,137 | 5,917 | -35% | 1 | 1 | 0% | 1,853 | 1,778 | -4% | 0 | 0 | — |
case-07 | fail→pass | 6,957 | 5,971 | -14% | 1 | 1 | 0% | 1,515 | 1,734 | +14% | 0 | 0 | — |
case-08 | fail→pass | 5,596 | 3,710 | -34% | 1 | 1 | 0% | 908 | 1,140 | +26% | 0 | 0 | — |
case-09 | fail→pass | 5,466 | 2,872 | -47% | 1 | 1 | 0% | 885 | 1,017 | +15% | 0 | 0 | — |
case-10 | pass→pass | 7,082 | 5,232 | -26% | 1 | 1 | 0% | 1,515 | 1,536 | +1% | 0 | 0 | — |
case-11 | fail→pass | 5,371 | 4,234 | -21% | 1 | 1 | 0% | 1,076 | 1,525 | +42% | 0 | 0 | — |
case-12 | fail→pass | 8,478 | 5,455 | -36% | 1 | 1 | 0% | 1,624 | 1,704 | +5% | 0 | 0 | — |
case-13 | pass→pass | 5,426 | 3,673 | -32% | 1 | 1 | 0% | 1,023 | 1,335 | +30% | 0 | 0 | — |
case-14 | pass→pass | 12,722 | 4,936 | -61% | 1 | 1 | 0% | 1,177 | 1,565 | +33% | 0 | 0 | — |
case-15 | pass→pass | 6,287 | 22,058 | +251% | 1 | 1 | 0% | 1,164 | 1,379 | +18% | 0 | 0 | — |
case-16 | pass→pass | 7,764 | 5,033 | -35% | 1 | 1 | 0% | 1,508 | 1,499 | -1% | 0 | 0 | — |
case-17 | pass→pass | 11,576 | 4,831 | -58% | 1 | 1 | 0% | 1,073 | 1,313 | +22% | 0 | 0 | — |
case-18 | fail→pass | 9,665 | 6,217 | -36% | 1 | 1 | 0% | 1,043 | 1,650 | +58% | 0 | 0 | — |
case-19 | fail→pass | 26,130 | 3,688 | -86% | 1 | 1 | 0% | 2,235 | 1,002 | -55% | 0 | 0 | — |
case-21 | fail→pass | 18,052 | 5,789 | -68% | 1 | 1 | 0% | 3,287 | 1,296 | -61% | 0 | 0 | — |
case-22 | fail→pass | 10,961 | 8,102 | -26% | 1 | 1 | 0% | 2,091 | 1,885 | -10% | 0 | 0 | — |
case-23 | fail→pass | 10,893 | 3,375 | -69% | 1 | 1 | 0% | 1,694 | 1,070 | -37% | 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. 23 cases were attempted. The headline lift of +61 percentage points is the difference between those two pass rates over the 23 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.