Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate NSIS installer scripts for Windows with custom UI and features
.claude/skills/a5c-ai-nsis-installer-generator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 60% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -1% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 24% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 14% | 0% |
Generate NSIS (Nullsoft Scriptable Install System) installer scripts for Windows applications.
json{ "type": "object", "properties": { "projectPath": { "type": "string" }, "appName": { "type": "string" }, "version": { "type": "string" }, "components": { "type": "array" }, "createShortcuts": { "type": "boolean" } }, "required": ["projectPath", "appName", "version"] }
nsi!include "MUI2.nsh" Name "My Application" OutFile "MyApp-Setup.exe" InstallDir "$PROGRAMFILES\MyApp" RequestExecutionLevel admin !insertmacro MUI_PAGE_WELCOME !insertmacro MUI_PAGE_DIRECTORY !insertmacro MUI_PAGE_INSTFILES !insertmacro MUI_PAGE_FINISH !insertmacro MUI_LANGUAGE "English" Section "Install" SetOutPath $INSTDIR File /r "dist\*.*" WriteUninstaller "$INSTDIR\Uninstall.exe" CreateShortcut "$DESKTOP\MyApp.lnk" "$INSTDIR\MyApp.exe" SectionEnd Section "Uninstall" Delete "$DESKTOP\MyApp.lnk" RMDir /r "$INSTDIR" SectionEnd
wix-toolset-configwindows-authenticode-signer| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 10,096 | 7,677 | -24% | 1 | 1 | 0% | 2,235 | 2,218 | -1% | 0 | 0 | — |
case-02 | pass→pass | 10,014 | 9,819 | -2% | 1 | 1 | 0% | 1,936 | 2,406 | +24% | 0 | 0 | — |
case-03 | pass→pass | 7,654 | 6,113 | -20% | 1 | 1 | 0% | 1,423 | 1,626 | +14% | 0 | 0 | — |
case-04 | pass→pass | 3,836 | 2,995 | -22% | 1 | 1 | 0% | 661 | 878 | +33% | 0 | 0 | — |
case-05 | pass→pass | 3,707 | 2,916 | -21% | 1 | 1 | 0% | 604 | 897 | +49% | 0 | 0 | — |
case-06 | pass→pass | 4,417 | 2,696 | -39% | 1 | 1 | 0% | 699 | 854 | +22% | 0 | 0 | — |
case-07 | pass→pass | 4,062 | 3,012 | -26% | 1 | 1 | 0% | 720 | 867 | +20% | 0 | 0 | — |
case-08 | pass→pass | 5,742 | 4,469 | -22% | 1 | 1 | 0% | 1,005 | 1,132 | +13% | 0 | 0 | — |
case-09 | pass→pass | 4,340 | 4,833 | +11% | 1 | 1 | 0% | 673 | 1,283 | +91% | 0 | 0 | — |
case-10 | pass→pass | 6,677 | 3,803 | -43% | 1 | 1 | 0% | 1,179 | 921 | -22% | 0 | 0 | — |
case-11 | pass→pass | 3,517 | 2,912 | -17% | 1 | 1 | 0% | 442 | 867 | +96% | 0 | 0 | — |
case-12 | pass→pass | 9,514 | 4,899 | -49% | 1 | 1 | 0% | 1,386 | 1,299 | -6% | 0 | 0 | — |
case-13 | pass→pass | 10,209 | 6,587 | -35% | 1 | 1 | 0% | 2,088 | 1,723 | -17% | 0 | 0 | — |
case-14 | pass→pass | 3,759 | 3,048 | -19% | 1 | 1 | 0% | 639 | 896 | +40% | 0 | 0 | — |
case-15 | pass→pass | 2,619 | 2,210 | -16% | 1 | 1 | 0% | 402 | 732 | +82% | 0 | 0 | — |
case-16 | pass→pass | 2,031 | 2,294 | +13% | 1 | 1 | 0% | 314 | 737 | +135% | 0 | 0 | — |
case-17 | fail→pass | 3,988 | 3,987 | -0% | 1 | 1 | 0% | 690 | 1,037 | +50% | 0 | 0 | — |
case-18 | fail→pass | 3,775 | 3,885 | +3% | 1 | 1 | 0% | 589 | 940 | +60% | 0 | 0 | — |
case-19 | pass→pass | 4,674 | 4,432 | -5% | 1 | 1 | 0% | 863 | 1,219 | +41% | 0 | 0 | — |
case-20 | pass→pass | 3,448 | 4,581 | +33% | 1 | 1 | 0% | 506 | 1,262 | +149% | 0 | 0 | — |
case-21 | pass→pass | 9,210 | 5,916 | -36% | 1 | 1 | 0% | 1,846 | 1,592 | -14% | 0 | 0 | — |
case-22 | pass→pass | 6,087 | 6,473 | +6% | 1 | 1 | 0% | 1,150 | 1,201 | +4% | 0 | 0 | — |
case-23 | pass→pass | 2,991 | 2,831 | -5% | 1 | 1 | 0% | 510 | 770 | +51% | 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 +9 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.