Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Implement rich text editors using Plate.js. Supports creating both simple (comment/chat) and detailed (document/blog) editors.
.claude/skills/aiskillstore-plate-handler/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -47% | 0% |
Use the shadcn CLI to install Plate components.
pnpm dlx shadcn@latest add @plate/editorpnpm dlx shadcn@latest add @plate/basic-nodes-kit @plate/fixed-toolbar @plate/mark-toolbar-buttonpnpm dlx shadcn@latest add @plate/editor-basic.Decide based on the user's requirement:
FixedToolbar or no toolbar (shortcuts only).src/components/plate-editor/simple-editor.tsxFixedToolbar with multiple groups, Floating Toolbar.src/components/plate-editor/detailed-editor.tsxEditor, EditorContainer, Toolbar) are installed in @/components/ui.usePlateEditor and <Plate>.plugins array.FixedToolbar and buttons (MarkToolbarButton, ToolbarButton) inside the <Plate> provider.value and onChange props on <Plate> to handle content. Sync with localStorage or form state as needed.See reference.md for code snippets, CLI commands, and configuration details.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-22 | pass→pass | 11,876 | 10,887 | -8% | 1 | 1 | 0% | 2,088 | 2,588 | +24% | 0 | 0 | — |
case-01 | fail→fail | 17,491 | 11,597 | -34% | 1 | 1 | 0% | 3,289 | 2,886 | -12% | 0 | 0 | — |
case-02 | fail→pass | 18,410 | 14,673 | -20% | 1 | 1 | 0% | 3,528 | 3,518 | -0% | 0 | 0 | — |
case-03 | fail→pass | 15,330 | 14,280 | -7% | 1 | 1 | 0% | 2,960 | 3,576 | +21% | 0 | 0 | — |
case-04 | fail→pass | 6,687 | 3,184 | -52% | 1 | 1 | 0% | 1,141 | 974 | -15% | 0 | 0 | — |
case-05 | pass→pass | 10,158 | 3,122 | -69% | 1 | 1 | 0% | 1,734 | 1,067 | -38% | 0 | 0 | — |
case-11 | fail→pass | 6,662 | 2,542 | -62% | 1 | 1 | 0% | 1,053 | 900 | -15% | 0 | 0 | — |
case-06 | fail→pass | 9,496 | 2,311 | -76% | 1 | 1 | 0% | 1,598 | 854 | -47% | 0 | 0 | — |
case-07 | fail→pass | 9,731 | 1,944 | -80% | 1 | 1 | 0% | 1,630 | 786 | -52% | 0 | 0 | — |
case-08 | fail→pass | 10,373 | 1,942 | -81% | 1 | 1 | 0% | 1,771 | 773 | -56% | 0 | 0 | — |
case-09 | pass→pass | 13,536 | 10,415 | -23% | 1 | 1 | 0% | 2,357 | 2,403 | +2% | 0 | 0 | — |
case-10 | fail→pass | 7,285 | 2,150 | -70% | 1 | 1 | 0% | 1,186 | 779 | -34% | 0 | 0 | — |
case-12 | fail→pass | 9,093 | 2,798 | -69% | 1 | 1 | 0% | 1,374 | 947 | -31% | 0 | 0 | — |
case-13 | pass→pass | 4,071 | 3,573 | -12% | 1 | 1 | 0% | 666 | 1,186 | +78% | 0 | 0 | — |
case-14 | fail→pass | 6,759 | 1,786 | -74% | 1 | 1 | 0% | 1,135 | 749 | -34% | 0 | 0 | — |
case-15 | pass→pass | 10,947 | 2,384 | -78% | 1 | 1 | 0% | 1,667 | 817 | -51% | 0 | 0 | — |
case-21 | pass→pass | 10,028 | 10,055 | +0% | 1 | 1 | 0% | 1,921 | 2,331 | +21% | 0 | 0 | — |
case-16 | pass→pass | 14,744 | 6,668 | -55% | 1 | 1 | 0% | 2,365 | 1,622 | -31% | 0 | 0 | — |
case-17 | fail→pass | 10,769 | 2,238 | -79% | 1 | 1 | 0% | 1,776 | 845 | -52% | 0 | 0 | — |
case-18 | pass→pass | 14,898 | 11,499 | -23% | 1 | 1 | 0% | 2,730 | 2,607 | -5% | 0 | 0 | — |
case-19 | fail→pass | 7,459 | 2,161 | -71% | 1 | 1 | 0% | 1,318 | 834 | -37% | 0 | 0 | — |
case-20 | pass→pass | 12,555 | 10,624 | -15% | 1 | 1 | 0% | 2,234 | 2,460 | +10% | 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 +55 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.