Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Add Agentation visual feedback toolbar to a Next.js project
.claude/skills/aiskillstore-agentation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -46% | 0% |
Set up the Agentation annotation toolbar in this project.
agentation in package.json dependenciesnpm install agentation (or pnpm/yarn based on lockfile)<Agentation or import { Agentation } in src/ or app/app/layout.tsx or app/layout.jspages/_app.tsx or pages/_app.jsFor Next.js App Router, add to the root layout: tsx import { Agentation } from "agentation";
// Add inside the body, after children: {process.env.NODE_ENV === "development" && <Agentation />}
For Next.js Pages Router, add to _app: tsx import { Agentation } from "agentation";
// Add after Component: {process.env.NODE_ENV === "development" && <Agentation />}
See add-mcp — run npx add-mcp and follow the prompts to add agentation-mcp as an MCP server
Run agentation-mcp init after installing the package
NODE_ENV check ensures Agentation only loads in developmentagentation_get_all_pending, agentation_resolve, and agentation_watch_annotationsagentation-mcp doctor to verify setup after installing| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 12,861 | 10,139 | -21% | 1 | 1 | 0% | 1,364 | 1,158 | -15% | 0 | 0 | — |
case-01 | fail→fail | 15,680 | 11,237 | -28% | 1 | 1 | 0% | 260 | 898 | +245% | 0 | 0 | — |
case-02 | fail→fail | 11,755 | 12,396 | +5% | 1 | 1 | 0% | 188 | 820 | +336% | 0 | 0 | — |
case-03 | fail→fail | 12,758 | 10,016 | -21% | 1 | 1 | 0% | 2,414 | 754 | -69% | 0 | 0 | — |
case-04 | pass→fail | 22,307 | 10,763 | -52% | 1 | 1 | 0% | 2,857 | 975 | -66% | 0 | 0 | — |
case-05 | pass→pass | 19,885 | 18,734 | -6% | 1 | 1 | 0% | 2,534 | 2,496 | -1% | 0 | 0 | — |
case-06 | pass→fail | 12,699 | 8,163 | -36% | 1 | 1 | 0% | 996 | 946 | -5% | 0 | 0 | — |
case-08 | fail→pass | 22,745 | 9,024 | -60% | 1 | 1 | 0% | 2,161 | 1,201 | -44% | 0 | 0 | — |
case-09 | pass→pass | 13,914 | 22,635 | +63% | 1 | 1 | 0% | 1,418 | 1,001 | -29% | 0 | 0 | — |
case-10 | pass→pass | 15,638 | 21,153 | +35% | 1 | 1 | 0% | 1,844 | 1,656 | -10% | 0 | 0 | — |
case-11 | pass→fail | 15,680 | 23,582 | +50% | 1 | 1 | 0% | 1,996 | 891 | -55% | 0 | 0 | — |
case-12 | fail→pass | 19,150 | 33,598 | +75% | 1 | 1 | 0% | 2,388 | 1,617 | -32% | 0 | 0 | — |
case-13 | fail→pass | 21,626 | 7,873 | -64% | 1 | 1 | 0% | 2,146 | 1,056 | -51% | 0 | 0 | — |
case-14 | pass→pass | 32,296 | 34,157 | +6% | 1 | 1 | 0% | 2,116 | 1,849 | -13% | 0 | 0 | — |
case-15 | fail→pass | 16,661 | 2,278 | -86% | 1 | 1 | 0% | 1,625 | 873 | -46% | 0 | 0 | — |
case-16 | fail→pass | 19,230 | 7,022 | -63% | 1 | 1 | 0% | 2,030 | 886 | -56% | 0 | 0 | — |
case-17 | pass→pass | 14,718 | 10,993 | -25% | 1 | 1 | 0% | 2,310 | 1,448 | -37% | 0 | 0 | — |
case-18 | fail→pass | 13,458 | 24,619 | +83% | 1 | 1 | 0% | 1,522 | 965 | -37% | 0 | 0 | — |
case-19 | pass→pass | 17,171 | 33,075 | +93% | 1 | 1 | 0% | 2,091 | 1,873 | -10% | 0 | 0 | — |
case-20 | pass→pass | 15,549 | 23,105 | +49% | 1 | 1 | 0% | 1,110 | 981 | -12% | 0 | 0 | — |
case-21 | pass→pass | 14,919 | 21,435 | +44% | 1 | 1 | 0% | 1,955 | 1,703 | -13% | 0 | 0 | — |
case-22 | fail→pass | 24,728 | 18,087 | -27% | 1 | 1 | 0% | 1,853 | 1,927 | +4% | 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, and 16 counted toward the lift figure. The other 6 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +23 percentage points is the difference between those two pass rates over the 16 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
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.