Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →You are a code migration expert specializing in transitioning codebases between frameworks, languages, versions, and platforms. Generate comprehensive migration plans, automated migration scripts, and
.claude/skills/dokhacgiakhoa-framework-migration-code-migrate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 50% | 0% |
You are a code migration expert specializing in transitioning codebases between frameworks, languages, versions, and platforms. Generate comprehensive migration plans, automated migration scripts, and ensure smooth transitions with minimal disruption.
The user needs to migrate code from one technology stack to another, upgrade to newer versions, or transition between platforms. Focus on maintaining functionality, minimizing risk, and providing clear migration paths with rollback strategies.
$ARGUMENTS
resources/implementation-playbook.md.Focus on minimizing disruption, maintaining functionality, and providing clear paths for successful code migration with comprehensive testing and rollback strategies.
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-12 | fail→pass | 22,287 | 20,352 | -9% | 1 | 1 | 0% | 3,462 | 3,937 | +14% | 0 | 0 | — |
case-01 | fail→fail | 55,618 | 57,180 | +3% | 1 | 1 | 0% | 8,295 | 8,683 | +5% | 0 | 0 | — |
case-02 | fail→fail | 55,496 | 54,124 | -2% | 1 | 1 | 0% | 8,278 | 8,666 | +5% | 0 | 0 | — |
case-03 | fail→fail | 46,713 | 51,678 | +11% | 1 | 1 | 0% | 8,288 | 8,676 | +5% | 0 | 0 | — |
case-04 | pass→pass | 27,809 | 33,033 | +19% | 1 | 1 | 0% | 4,406 | 5,152 | +17% | 0 | 0 | — |
case-05 | fail→fail | 7,936 | 24,373 | +207% | 1 | 1 | 0% | 1,110 | 4,900 | +341% | 0 | 0 | — |
case-06 | pass→fail | 18,389 | 61,789 | +236% | 1 | 1 | 0% | 2,965 | 1,637 | -45% | 0 | 0 | — |
case-07 | fail→pass | 49,405 | 45,099 | -9% | 1 | 1 | 0% | 8,247 | 8,635 | +5% | 0 | 0 | — |
case-08 | fail→pass | 51,077 | 49,645 | -3% | 1 | 1 | 0% | 8,237 | 7,899 | -4% | 0 | 0 | — |
case-09 | pass→pass | 45,930 | 36,652 | -20% | 1 | 1 | 0% | 7,332 | 8,630 | +18% | 0 | 0 | — |
case-10 | fail→pass | 36,990 | 34,659 | -6% | 1 | 1 | 0% | 6,740 | 6,502 | -4% | 0 | 0 | — |
case-11 | fail→pass | 28,533 | 37,380 | +31% | 1 | 1 | 0% | 5,104 | 7,654 | +50% | 0 | 0 | — |
case-13 | pass→pass | 49,932 | 40,650 | -19% | 1 | 1 | 0% | 8,218 | 8,606 | +5% | 0 | 0 | — |
case-14 | pass→pass | 22,739 | 37,352 | +64% | 1 | 1 | 0% | 3,998 | 6,962 | +74% | 0 | 0 | — |
case-15 | fail→pass | 22,968 | 36,016 | +57% | 1 | 1 | 0% | 4,692 | 8,352 | +78% | 0 | 0 | — |
case-16 | fail→pass | 46,371 | 44,974 | -3% | 1 | 1 | 0% | 8,224 | 8,612 | +5% | 0 | 0 | — |
case-17 | pass→pass | 40,104 | 52,135 | +30% | 1 | 1 | 0% | 5,044 | 8,360 | +66% | 0 | 0 | — |
case-18 | fail→pass | 21,766 | 39,963 | +84% | 1 | 1 | 0% | 3,632 | 6,539 | +80% | 0 | 0 | — |
case-19 | pass→pass | 23,131 | 35,278 | +53% | 1 | 1 | 0% | 4,148 | 6,683 | +61% | 0 | 0 | — |
case-20 | fail→fail | 27,962 | 34,628 | +24% | 1 | 1 | 0% | 4,216 | 6,767 | +61% | 0 | 0 | — |
case-21 | fail→fail | 29,373 | 44,040 | +50% | 1 | 1 | 0% | 5,279 | 8,561 | +62% | 0 | 0 | — |
case-22 | fail→fail | 20,387 | 28,870 | +42% | 1 | 1 | 0% | 3,435 | 5,872 | +71% | 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 21 counted toward the lift figure. The other 1 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 +32 percentage points is the difference between those two pass rates over the 21 comparable cases. 1 case got worse with the skill loaded, and it is 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.