Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When building applications on Cloudflare Workers or with frameworks that deploy to Cloudflare Workers, strongly prefer retrieval and fetching documentation over training data.
.claude/skills/kunanonj-cursor-plugin-cf-rule-workers/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-17 | ✓→✗ | ▼ Worse | 19% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | -5% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -41% | 0% |
STOP. Your knowledge of Cloudflare Workers APIs and limits may be outdated. Always retrieve current documentation before any Workers, KV, R2, D1, Durable Objects, Queues, Vectorize, AI, or Agents SDK task.
https://docs.mcp.cloudflare.com/mcpFor all limits and quotas, retrieve from the product's /platform/limits/ page.
| Command | Purpose | |---------|---------| | npx wrangler dev | Local development | | npx wrangler deploy | Deploy to Cloudflare | | npx wrangler types | Generate TypeScript types |
Run wrangler types after changing bindings in wrangler.jsonc.
If you encounter Dynamic require of "X" is not supported or missing Node.js APIs:
jsonc{ "compatibility_flags": ["nodejs_compat"], "compatibility_date": "YYYY-MM-DD" // Use today's date }
Docs: https://developers.cloudflare.com/workers/runtime-apis/nodejs/
/workers/platform/limits/Retrieve API references and limits from: /kv/ · /r2/ · /d1/ · /durable-objects/ · /queues/ · /vectorize/ · /workers-ai/ · /agents/
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-22 | pass→pass | 12,107 | 9,119 | -25% | 1 | 1 | 0% | 2,553 | 2,432 | -5% | 0 | 0 | — |
case-02 | fail→fail | 10,750 | 4,913 | -54% | 1 | 1 | 0% | 2,303 | 1,561 | -32% | 0 | 0 | — |
case-04 | pass→pass | 5,059 | 1,500 | -70% | 1 | 1 | 0% | 1,159 | 679 | -41% | 0 | 0 | — |
case-03 | pass→pass | 2,491 | 1,658 | -33% | 1 | 1 | 0% | 528 | 743 | +41% | 0 | 0 | — |
case-01 | fail→pass | 11,574 | 9,991 | -14% | 1 | 1 | 0% | 2,303 | 2,537 | +10% | 0 | 0 | — |
case-05 | pass→pass | 3,779 | 2,569 | -32% | 1 | 1 | 0% | 823 | 861 | +5% | 0 | 0 | — |
case-06 | pass→pass | 9,396 | 2,834 | -70% | 1 | 1 | 0% | 1,949 | 994 | -49% | 0 | 0 | — |
case-07 | pass→pass | 6,627 | 2,990 | -55% | 1 | 1 | 0% | 1,447 | 1,043 | -28% | 0 | 0 | — |
case-08 | pass→pass | 5,121 | 2,907 | -43% | 1 | 1 | 0% | 1,024 | 979 | -4% | 0 | 0 | — |
case-09 | pass→pass | 5,868 | 2,606 | -56% | 1 | 1 | 0% | 1,180 | 977 | -17% | 0 | 0 | — |
case-10 | pass→pass | 3,382 | 1,742 | -48% | 1 | 1 | 0% | 616 | 735 | +19% | 0 | 0 | — |
case-11 | pass→pass | 7,993 | 2,527 | -68% | 1 | 1 | 0% | 1,490 | 961 | -36% | 0 | 0 | — |
case-12 | pass→pass | 2,399 | 1,435 | -40% | 1 | 1 | 0% | 468 | 682 | +46% | 0 | 0 | — |
case-13 | pass→pass | 6,531 | 4,211 | -36% | 1 | 1 | 0% | 1,469 | 1,331 | -9% | 0 | 0 | — |
case-14 | pass→pass | 4,331 | 2,412 | -44% | 1 | 1 | 0% | 815 | 719 | -12% | 0 | 0 | — |
case-15 | fail→pass | 9,116 | 1,974 | -78% | 1 | 1 | 0% | 1,965 | 797 | -59% | 0 | 0 | — |
case-16 | pass→pass | 13,183 | 7,027 | -47% | 1 | 1 | 0% | 2,744 | 1,844 | -33% | 0 | 0 | — |
case-17 | pass→fail | 3,404 | 1,913 | -44% | 1 | 1 | 0% | 637 | 760 | +19% | 0 | 0 | — |
case-18 | pass→pass | 5,482 | 2,752 | -50% | 1 | 1 | 0% | 1,222 | 900 | -26% | 0 | 0 | — |
case-19 | pass→pass | 1,966 | 1,734 | -12% | 1 | 1 | 0% | 287 | 621 | +116% | 0 | 0 | — |
case-20 | pass→pass | 3,105 | 3,497 | +13% | 1 | 1 | 0% | 653 | 1,106 | +69% | 0 | 0 | — |
case-21 | fail→fail | 2,811 | 2,466 | -12% | 1 | 1 | 0% | 475 | 890 | +87% | 0 | 0 | — |
case-23 | pass→pass | 4,529 | 3,608 | -20% | 1 | 1 | 0% | 950 | 1,144 | +20% | 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 +4 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is 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.