Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create SQLx migration files with `sqlx migrate add <name>`. Use when asked to add, create, or generate a sqlx/sqlx-cli database migration.
.claude/skills/macro-inc-sqlx-migration/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 56% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -27% | 0% |
Use the SQLx CLI to create migration files instead of manually creating timestamped files.
migrations/ directory.create_users_table.just setup_test_envs from the project rootcrates/macro_db_client folder:bash cargo sqlx migrate add name_of_migration
Replace name_of_migration with the actual snake_case migration name.
sqlx migrate add generates the correct timestamped file.sqlx is not installed or not on PATH, stop and tell the user to install sqlx-cli rather than manually creating the migration file.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 9,981 | 6,662 | -33% | 1 | 1 | 0% | 1,866 | 559 | -70% | 0 | 0 | — |
case-02 | fail→fail | 8,673 | 13,051 | +50% | 1 | 1 | 0% | 1,469 | 439 | -70% | 0 | 0 | — |
case-03 | fail→fail | 9,107 | 6,382 | -30% | 1 | 1 | 0% | 1,442 | 614 | -57% | 0 | 0 | — |
case-04 | fail→pass | 13,871 | 6,108 | -56% | 1 | 1 | 0% | 2,142 | 1,238 | -42% | 0 | 0 | — |
case-05 | fail→pass | 8,865 | 16,258 | +83% | 1 | 1 | 0% | 1,332 | 2,072 | +56% | 0 | 0 | — |
case-06 | fail→pass | 11,602 | 4,569 | -61% | 1 | 1 | 0% | 1,958 | 959 | -51% | 0 | 0 | — |
case-07 | fail→pass | 11,862 | 2,866 | -76% | 1 | 1 | 0% | 1,047 | 708 | -32% | 0 | 0 | — |
case-08 | fail→fail | 13,451 | 10,749 | -20% | 1 | 1 | 0% | 2,272 | 776 | -66% | 0 | 0 | — |
case-09 | fail→fail | 6,709 | 3,231 | -52% | 1 | 1 | 0% | 1,123 | 722 | -36% | 0 | 0 | — |
case-10 | pass→pass | 7,398 | 3,794 | -49% | 1 | 1 | 0% | 1,058 | 762 | -28% | 0 | 0 | — |
case-11 | pass→pass | 9,609 | 6,946 | -28% | 1 | 1 | 0% | 1,661 | 1,421 | -14% | 0 | 0 | — |
case-12 | fail→pass | 11,088 | 8,396 | -24% | 1 | 1 | 0% | 1,842 | 1,338 | -27% | 0 | 0 | — |
case-13 | fail→pass | 11,194 | 3,613 | -68% | 1 | 1 | 0% | 1,600 | 737 | -54% | 0 | 0 | — |
case-14 | fail→pass | 10,795 | 3,381 | -69% | 1 | 1 | 0% | 1,747 | 747 | -57% | 0 | 0 | — |
case-15 | pass→pass | 14,943 | 5,378 | -64% | 1 | 1 | 0% | 1,155 | 1,038 | -10% | 0 | 0 | — |
case-16 | fail→pass | 15,912 | 126,109 | +693% | 1 | 1 | 0% | 1,781 | 800 | -55% | 0 | 0 | — |
case-17 | fail→fail | 10,334 | 4,203 | -59% | 1 | 1 | 0% | 1,340 | 818 | -39% | 0 | 0 | — |
case-18 | fail→pass | 9,923 | 5,234 | -47% | 1 | 1 | 0% | 1,645 | 1,035 | -37% | 0 | 0 | — |
case-19 | fail→pass | 12,789 | 3,328 | -74% | 1 | 1 | 0% | 1,811 | 706 | -61% | 0 | 0 | — |
case-20 | pass→pass | 9,067 | 6,848 | -24% | 1 | 1 | 0% | 1,323 | 1,337 | +1% | 0 | 0 | — |
case-21 | pass→pass | 15,982 | 11,421 | -29% | 1 | 1 | 0% | 2,606 | 2,124 | -18% | 0 | 0 | — |
case-22 | pass→pass | 9,463 | 8,615 | -9% | 1 | 1 | 0% | 1,399 | 1,776 | +27% | 0 | 0 | — |
case-23 | pass→pass | 11,845 | 5,937 | -50% | 1 | 1 | 0% | 1,693 | 1,189 | -30% | 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, and 20 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 +43 percentage points is the difference between those two pass rates over the 20 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.