Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →You are a dependency management expert specializing in safe, incremental upgrades of project dependencies. Plan and execute dependency updates with minimal risk, proper testing, and clear migration pa
.claude/skills/dokhacgiakhoa-framework-migration-deps-upgrade/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 84% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 36% | 0% |
You are a dependency management expert specializing in safe, incremental upgrades of project dependencies. Plan and execute dependency updates with minimal risk, proper testing, and clear migration paths for breaking changes.
The user needs to upgrade project dependencies safely, handling breaking changes, ensuring compatibility, and maintaining stability. Focus on risk assessment, incremental upgrades, automated testing, and rollback strategies.
$ARGUMENTS
resources/implementation-playbook.md.Focus on safe, incremental upgrades that maintain system stability while keeping dependencies current and secure.
resources/implementation-playbook.md for detailed patterns and examples.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | pass→fail | 19,371 | 22,602 | +17% | 1 | 1 | 0% | 3,076 | 3,809 | +24% | 0 | 0 | — |
case-01 | pass→pass | 51,454 | 40,228 | -22% | 1 | 1 | 0% | 8,274 | 6,952 | -16% | 0 | 0 | — |
case-03 | pass→pass | 11,176 | 13,161 | +18% | 1 | 1 | 0% | 1,649 | 2,353 | +43% | 0 | 0 | — |
case-04 | pass→pass | 10,204 | 7,123 | -30% | 1 | 1 | 0% | 1,438 | 1,589 | +11% | 0 | 0 | — |
case-05 | fail→pass | 22,859 | 30,335 | +33% | 1 | 1 | 0% | 3,409 | 4,620 | +36% | 0 | 0 | — |
case-06 | fail→pass | 25,432 | 76,095 | +199% | 1 | 1 | 0% | 3,527 | 5,804 | +65% | 0 | 0 | — |
case-07 | pass→pass | 31,058 | 35,277 | +14% | 1 | 1 | 0% | 4,434 | 6,297 | +42% | 0 | 0 | — |
case-08 | fail→pass | 22,855 | 29,945 | +31% | 1 | 1 | 0% | 3,261 | 5,695 | +75% | 0 | 0 | — |
case-09 | fail→pass | 22,394 | 36,997 | +65% | 1 | 1 | 0% | 3,171 | 5,832 | +84% | 0 | 0 | — |
case-10 | fail→pass | 18,315 | 26,686 | +46% | 1 | 1 | 0% | 3,161 | 4,294 | +36% | 0 | 0 | — |
case-11 | pass→pass | 27,649 | 42,880 | +55% | 1 | 1 | 0% | 3,861 | 8,513 | +120% | 0 | 0 | — |
case-12 | fail→pass | 21,272 | 34,327 | +61% | 1 | 1 | 0% | 3,568 | 5,642 | +58% | 0 | 0 | — |
case-13 | fail→pass | 15,279 | 27,280 | +79% | 1 | 1 | 0% | 2,611 | 4,515 | +73% | 0 | 0 | — |
case-14 | fail→pass | 21,671 | 35,855 | +65% | 1 | 1 | 0% | 3,651 | 5,551 | +52% | 0 | 0 | — |
case-15 | pass→fail | 24,409 | 25,370 | +4% | 1 | 1 | 0% | 3,561 | 4,001 | +12% | 0 | 0 | — |
case-16 | fail→pass | 19,379 | 25,948 | +34% | 1 | 1 | 0% | 3,275 | 5,147 | +57% | 0 | 0 | — |
case-17 | fail→pass | 15,522 | 22,070 | +42% | 1 | 1 | 0% | 2,594 | 4,201 | +62% | 0 | 0 | — |
case-18 | fail→pass | 19,732 | 35,983 | +82% | 1 | 1 | 0% | 3,691 | 7,626 | +107% | 0 | 0 | — |
case-19 | fail→pass | 16,859 | 26,346 | +56% | 1 | 1 | 0% | 2,651 | 5,013 | +89% | 0 | 0 | — |
case-20 | fail→pass | 16,175 | 20,905 | +29% | 1 | 1 | 0% | 2,809 | 3,772 | +34% | 0 | 0 | — |
case-21 | fail→pass | 19,372 | 35,675 | +84% | 1 | 1 | 0% | 3,197 | 6,463 | +102% | 0 | 0 | — |
case-22 | fail→pass | 16,064 | 27,433 | +71% | 1 | 1 | 0% | 2,616 | 5,759 | +120% | 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 +59 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.