Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →二维码生成 - 把网址/文本/WiFi 信息生成二维码图片，本地离线生成
.claude/skills/dongsheng123132-qrcode-maker/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -84% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -83% | 0% |
| case-11 | ✓→✗ | ▼ Worse | -7% | 0% |
| case-12 | ✓→✗ | ▼ Worse | -88% | 0% |
帮用户把网址、文本、联系方式、WiFi 信息生成二维码图片，本地离线生成，不上传任何数据。
用 Bash 工具，Python 的 qrcode 库（轻量、纯本地）：
bashpython -c "import qrcode" 2>/dev/null || pip install -q "qrcode[pil]" # 网址/文本 -> PNG python - <<'PY' import qrcode qrcode.make("https://u-claw.org").save("qrcode.png") print("已生成 -> qrcode.png") PY # WiFi 二维码（扫码自动连网） python - <<'PY' import qrcode ssid, pwd, enc = "MyWiFi", "password123", "WPA" # enc: WPA / WEP / nopass data = f"WIFI:T:{enc};S:{ssid};P:{pwd};;" qrcode.make(data).save("wifi-qr.png") print("已生成 WiFi 二维码 -> wifi-qr.png") PY # 终端字符画预览（不存文件，直接看） python - <<'PY' import qrcode q = qrcode.QRCode(); q.add_data("https://u-claw.org"); q.make() q.print_ascii(invert=True) PY
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 13,063 | 15,062 | +15% | 1 | 1 | 0% | 1,767 | 1,788 | +1% | 0 | 0 | — |
case-01 | fail→fail | 8,085 | 16,906 | +109% | 1 | 1 | 0% | 1,412 | 1,163 | -18% | 0 | 0 | — |
case-02 | fail→fail | 50,198 | 11,338 | -77% | 1 | 1 | 0% | 8,232 | 965 | -88% | 0 | 0 | — |
case-03 | fail→fail | 9,881 | 11,024 | +12% | 1 | 1 | 0% | 1,664 | 1,256 | -25% | 0 | 0 | — |
case-04 | pass→fail | 29,455 | 12,073 | -59% | 1 | 1 | 0% | 5,335 | 851 | -84% | 0 | 0 | — |
case-05 | pass→fail | 22,039 | 12,629 | -43% | 1 | 1 | 0% | 5,289 | 925 | -83% | 0 | 0 | — |
case-07 | fail→fail | 8,738 | 8,909 | +2% | 1 | 1 | 0% | 1,178 | 773 | -34% | 0 | 0 | — |
case-08 | fail→fail | 12,821 | 7,914 | -38% | 1 | 1 | 0% | 2,474 | 888 | -64% | 0 | 0 | — |
case-09 | fail→pass | 7,556 | 5,088 | -33% | 1 | 1 | 0% | 1,394 | 1,092 | -22% | 0 | 0 | — |
case-10 | fail→fail | 10,496 | 35,180 | +235% | 1 | 1 | 0% | 2,062 | 1,059 | -49% | 0 | 0 | — |
case-11 | pass→fail | 5,937 | 13,301 | +124% | 1 | 1 | 0% | 865 | 808 | -7% | 0 | 0 | — |
case-12 | pass→fail | 31,419 | 7,420 | -76% | 1 | 1 | 0% | 5,459 | 653 | -88% | 0 | 0 | — |
case-13 | pass→pass | 10,207 | 14,891 | +46% | 1 | 1 | 0% | 1,467 | 2,529 | +72% | 0 | 0 | — |
case-14 | pass→fail | 8,376 | 12,657 | +51% | 1 | 1 | 0% | 1,480 | 808 | -45% | 0 | 0 | — |
case-15 | pass→fail | 8,197 | 12,139 | +48% | 1 | 1 | 0% | 1,128 | 1,221 | +8% | 0 | 0 | — |
case-16 | fail→fail | 44,465 | 6,836 | -85% | 1 | 1 | 0% | 8,215 | 766 | -91% | 0 | 0 | — |
case-17 | pass→fail | 8,252 | 8,794 | +7% | 1 | 1 | 0% | 1,697 | 999 | -41% | 0 | 0 | — |
case-18 | pass→fail | 5,467 | 11,079 | +103% | 1 | 1 | 0% | 1,060 | 1,412 | +33% | 0 | 0 | — |
case-19 | pass→fail | 39,732 | 7,014 | -82% | 1 | 1 | 0% | 7,210 | 739 | -90% | 0 | 0 | — |
case-20 | pass→pass | 12,438 | 11,822 | -5% | 1 | 1 | 0% | 1,751 | 2,038 | +16% | 0 | 0 | — |
case-21 | pass→fail | 11,791 | 7,976 | -32% | 1 | 1 | 0% | 2,114 | 751 | -64% | 0 | 0 | — |
case-22 | pass→fail | 8,975 | 19,266 | +115% | 1 | 1 | 0% | 1,432 | 1,314 | -8% | 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 5 counted toward the lift figure. The other 17 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 -45 percentage points is the difference between those two pass rates over the 5 comparable cases. 12 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.