Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when creating, linking, editing, importing/exporting, diagnosing, or version-recovering MaxCompute semantic profiles.
.claude/skills/aliyun-onboarding/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -81% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -72% | 0% |
Use this workflow for setup and profile maintenance, not for answering a single data question.
bashmcs -f json doctor mcs profile create mcs link bind PROFILE mcs profile show PROFILE --format json mcs profile update PROFILE --from-spec '<json>' mcs profile export PROFILE mcs profile import PATH
For profile history and recovery:
bashmcs profile log --profile PROFILE mcs profile log-show REF --profile PROFILE mcs profile diff REF_A REF_B --profile PROFILE mcs profile reset --to REF --profile PROFILE mcs profile fork FORK_NAME --from REF --profile PROFILE
See references/onboarding.md, references/profile-editor.md, and references/profile-history.md.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 9,365 | 3,168 | -66% | 1 | 1 | 0% | 1,563 | 749 | -52% | 0 | 0 | — |
case-02 | fail→pass | 16,664 | 2,193 | -87% | 1 | 1 | 0% | 2,719 | 508 | -81% | 0 | 0 | — |
case-03 | fail→pass | 12,575 | 3,494 | -72% | 1 | 1 | 0% | 2,016 | 821 | -59% | 0 | 0 | — |
case-04 | fail→pass | 6,461 | 1,723 | -73% | 1 | 1 | 0% | 986 | 385 | -61% | 0 | 0 | — |
case-05 | fail→pass | 9,328 | 1,635 | -82% | 1 | 1 | 0% | 1,562 | 439 | -72% | 0 | 0 | — |
case-06 | pass→pass | 9,609 | 1,413 | -85% | 1 | 1 | 0% | 1,584 | 353 | -78% | 0 | 0 | — |
case-07 | fail→pass | 17,386 | 2,524 | -85% | 1 | 1 | 0% | 1,912 | 622 | -67% | 0 | 0 | — |
case-08 | fail→pass | 10,084 | 2,034 | -80% | 1 | 1 | 0% | 1,748 | 537 | -69% | 0 | 0 | — |
case-09 | fail→pass | 10,369 | 2,586 | -75% | 1 | 1 | 0% | 1,590 | 652 | -59% | 0 | 0 | — |
case-10 | fail→pass | 12,528 | 1,785 | -86% | 1 | 1 | 0% | 2,069 | 364 | -82% | 0 | 0 | — |
case-11 | fail→pass | 9,960 | 1,963 | -80% | 1 | 1 | 0% | 1,575 | 448 | -72% | 0 | 0 | — |
case-12 | fail→pass | 14,368 | 1,898 | -87% | 1 | 1 | 0% | 2,211 | 442 | -80% | 0 | 0 | — |
case-13 | fail→pass | 12,347 | 2,060 | -83% | 1 | 1 | 0% | 1,986 | 399 | -80% | 0 | 0 | — |
case-14 | pass→pass | 7,372 | 2,570 | -65% | 1 | 1 | 0% | 1,197 | 589 | -51% | 0 | 0 | — |
case-15 | fail→pass | 9,066 | 1,838 | -80% | 1 | 1 | 0% | 1,515 | 453 | -70% | 0 | 0 | — |
case-16 | fail→pass | 14,620 | 1,848 | -87% | 1 | 1 | 0% | 2,192 | 451 | -79% | 0 | 0 | — |
case-17 | fail→pass | 5,431 | 2,302 | -58% | 1 | 1 | 0% | 787 | 502 | -36% | 0 | 0 | — |
case-18 | fail→pass | 14,324 | 2,148 | -85% | 1 | 1 | 0% | 2,348 | 480 | -80% | 0 | 0 | — |
case-19 | pass→pass | 9,565 | 1,499 | -84% | 1 | 1 | 0% | 1,559 | 357 | -77% | 0 | 0 | — |
case-20 | pass→pass | 6,451 | 5,086 | -21% | 1 | 1 | 0% | 1,213 | 1,010 | -17% | 0 | 0 | — |
case-21 | pass→pass | 6,975 | 5,083 | -27% | 1 | 1 | 0% | 1,310 | 1,058 | -19% | 0 | 0 | — |
case-22 | pass→pass | 5,621 | 4,266 | -24% | 1 | 1 | 0% | 1,184 | 937 | -21% | 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. The headline lift of +73 percentage points is the difference between those two pass rates over the 22 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.