Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build (compile) the solution via the developer CLI - backend (.NET), frontend (React/TypeScript), and the developer CLI itself.
.claude/skills/platformplatform-build/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -78% | 0% |
bashdotnet run --project developer-cli -- build [--backend] [--frontend] [--cli] [--self-contained-system <name>] --quiet
Use developer-cli exactly as written - do not expand to an absolute worktree path.
--backend - .NET--frontend - React/TypeScript--cli - the developer CLI itself--self-contained-system <name> - narrows the backend build to one SCS (e.g. account, main)No arguments builds everything.
bashdotnet run --project developer-cli -- build --quiet # everything dotnet run --project developer-cli -- build --backend --quiet # all backend dotnet run --project developer-cli -- build --frontend --quiet # frontend dotnet run --project developer-cli -- build --backend --self-contained-system main --quiet # one SCS
Verbose output goes to a log file. On success the CLI prints a single line; on failure it prints a short error summary - read the log if you need the full errors. Without --quiet the build floods the conversation.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-12 | fail→pass | 11,711 | 2,039 | -83% | 1 | 1 | 0% | 1,857 | 554 | -70% | 0 | 0 | — |
case-22 | pass→pass | 11,392 | 2,672 | -77% | 1 | 1 | 0% | 1,814 | 698 | -62% | 0 | 0 | — |
case-02 | fail→fail | 4,337 | 4,398 | +1% | 1 | 1 | 0% | 651 | 404 | -38% | 0 | 0 | — |
case-01 | fail→fail | 3,422 | 4,374 | +28% | 1 | 1 | 0% | 506 | 434 | -14% | 0 | 0 | — |
case-03 | fail→fail | 4,569 | 4,644 | +2% | 1 | 1 | 0% | 722 | 445 | -38% | 0 | 0 | — |
case-04 | fail→pass | 10,916 | 3,794 | -65% | 1 | 1 | 0% | 1,681 | 779 | -54% | 0 | 0 | — |
case-05 | fail→pass | 9,154 | 1,711 | -81% | 1 | 1 | 0% | 1,503 | 518 | -66% | 0 | 0 | — |
case-06 | fail→pass | 11,965 | 2,437 | -80% | 1 | 1 | 0% | 1,813 | 613 | -66% | 0 | 0 | — |
case-07 | fail→pass | 12,184 | 4,151 | -66% | 1 | 1 | 0% | 1,950 | 428 | -78% | 0 | 0 | — |
case-08 | pass→pass | 13,135 | 2,340 | -82% | 1 | 1 | 0% | 2,103 | 634 | -70% | 0 | 0 | — |
case-09 | pass→pass | 8,231 | 1,749 | -79% | 1 | 1 | 0% | 1,330 | 519 | -61% | 0 | 0 | — |
case-10 | pass→pass | 11,093 | 2,696 | -76% | 1 | 1 | 0% | 1,873 | 632 | -66% | 0 | 0 | — |
case-11 | fail→pass | 8,452 | 2,121 | -75% | 1 | 1 | 0% | 1,341 | 538 | -60% | 0 | 0 | — |
case-13 | pass→pass | 5,290 | 1,857 | -65% | 1 | 1 | 0% | 797 | 488 | -39% | 0 | 0 | — |
case-14 | fail→pass | 11,120 | 1,444 | -87% | 1 | 1 | 0% | 1,623 | 457 | -72% | 0 | 0 | — |
case-15 | fail→pass | 11,025 | 1,355 | -88% | 1 | 1 | 0% | 1,809 | 403 | -78% | 0 | 0 | — |
case-16 | fail→pass | 11,960 | 1,206 | -90% | 1 | 1 | 0% | 2,185 | 363 | -83% | 0 | 0 | — |
case-17 | pass→pass | 9,310 | 1,980 | -79% | 1 | 1 | 0% | 1,505 | 528 | -65% | 0 | 0 | — |
case-18 | fail→fail | 4,074 | 2,747 | -33% | 1 | 1 | 0% | 575 | 549 | -5% | 0 | 0 | — |
case-19 | fail→pass | 10,160 | 7,862 | -23% | 1 | 1 | 0% | 1,568 | 1,059 | -32% | 0 | 0 | — |
case-20 | fail→pass | 8,897 | 3,036 | -66% | 1 | 1 | 0% | 1,432 | 692 | -52% | 0 | 0 | — |
case-21 | fail→pass | 11,437 | 2,151 | -81% | 1 | 1 | 0% | 1,885 | 601 | -68% | 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 19 counted toward the lift figure. The other 3 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 +55 percentage points is the difference between those two pass rates over the 19 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.