Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Implementation details for EF Core migrations. Use when changing MigrationsSqlGenerator, model diffing, migration operations, HistoryRepository, the Migrator or related classes.
.claude/skills/migrations/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -60% | 0% |
Add migration: MigrationsScaffolder.ScaffoldMigration() → MigrationsModelDiffer.GetDifferences() → list of MigrationOperation → CSharpMigrationsGenerator and CSharpSnapshotGenerator produce Up/Down/Snapshot code
Apply migration: Migrator.MigrateAsync() → reads __EFMigrationsHistory → per pending: MigrationsSqlGenerator.Generate(operations) → MigrationCommandExecutor executes
typeof(Dictionary<string, object>) (property bag format), not the actual CLR type. When examining the ClrType in a snapshot, don't assume it matches the real entity type.SnapshotModelProcessor.Process() is used at design-time to fixup older model snapshots for backward compatibility.Migration operation tests: test/EFCore.Relational.Tests/Migrations/. Functional tests: test/EFCore.{Provider}.FunctionalTests/Migrations/. Model differ tests: test/EFCore.Relational.Tests/Migrations/Internal/MigrationsModelDifferTest*.cs.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 14,395 | 13,492 | -6% | 1 | 1 | 0% | 2,585 | 2,803 | +8% | 0 | 0 | — |
case-02 | fail→pass | 18,332 | 12,351 | -33% | 1 | 1 | 0% | 3,142 | 2,398 | -24% | 0 | 0 | — |
case-03 | fail→pass | 9,942 | 3,462 | -65% | 1 | 1 | 0% | 1,800 | 906 | -50% | 0 | 0 | — |
case-04 | fail→fail | 10,828 | 3,762 | -65% | 1 | 1 | 0% | 2,097 | 972 | -54% | 0 | 0 | — |
case-05 | fail→fail | 7,343 | 2,957 | -60% | 1 | 1 | 0% | 1,470 | 793 | -46% | 0 | 0 | — |
case-06 | fail→pass | 6,480 | 3,954 | -39% | 1 | 1 | 0% | 1,277 | 915 | -28% | 0 | 0 | — |
case-07 | fail→fail | 6,292 | 3,908 | -38% | 1 | 1 | 0% | 1,215 | 936 | -23% | 0 | 0 | — |
case-08 | fail→fail | 7,107 | 4,213 | -41% | 1 | 1 | 0% | 1,321 | 948 | -28% | 0 | 0 | — |
case-09 | fail→pass | 11,989 | 3,900 | -67% | 1 | 1 | 0% | 2,235 | 903 | -60% | 0 | 0 | — |
case-10 | fail→fail | 14,040 | 5,308 | -62% | 1 | 1 | 0% | 2,310 | 1,074 | -54% | 0 | 0 | — |
case-11 | pass→pass | 12,356 | 2,152 | -83% | 1 | 1 | 0% | 2,212 | 628 | -72% | 0 | 0 | — |
case-12 | fail→pass | 9,085 | 2,135 | -76% | 1 | 1 | 0% | 1,838 | 628 | -66% | 0 | 0 | — |
case-13 | fail→fail | 5,657 | 1,947 | -66% | 1 | 1 | 0% | 1,088 | 595 | -45% | 0 | 0 | — |
case-14 | fail→fail | 10,564 | 2,346 | -78% | 1 | 1 | 0% | 2,016 | 605 | -70% | 0 | 0 | — |
case-15 | fail→fail | 11,482 | 2,059 | -82% | 1 | 1 | 0% | 1,016 | 554 | -45% | 0 | 0 | — |
case-16 | fail→fail | 10,488 | 8,460 | -19% | 1 | 1 | 0% | 2,154 | 1,913 | -11% | 0 | 0 | — |
case-17 | fail→pass | 9,400 | 4,109 | -56% | 1 | 1 | 0% | 1,640 | 931 | -43% | 0 | 0 | — |
case-18 | fail→fail | 14,711 | 6,340 | -57% | 1 | 1 | 0% | 2,822 | 1,517 | -46% | 0 | 0 | — |
case-19 | fail→fail | 9,303 | 7,957 | -14% | 1 | 1 | 0% | 1,790 | 1,782 | -0% | 0 | 0 | — |
case-20 | fail→fail | 5,295 | 3,159 | -40% | 1 | 1 | 0% | 1,032 | 829 | -20% | 0 | 0 | — |
case-21 | fail→fail | 8,667 | 6,309 | -27% | 1 | 1 | 0% | 1,642 | 1,549 | -6% | 0 | 0 | — |
case-22 | fail→fail | 7,418 | 5,968 | -20% | 1 | 1 | 0% | 1,669 | 1,400 | -16% | 0 | 0 | — |
case-23 | fail→fail | 10,563 | 10,012 | -5% | 1 | 1 | 0% | 2,251 | 2,124 | -6% | 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 +30 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.