Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Orchestrate a comprehensive legacy system modernization using the strangler fig pattern, enabling gradual replacement of outdated components while maintaining continuous business operations through ex
.claude/skills/dokhacgiakhoa-framework-migration-legacy-modernize/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | -60% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 27% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 30% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | -24% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | 2% | 0% |
Orchestrate a comprehensive legacy system modernization using the strangler fig pattern, enabling gradual replacement of outdated components while maintaining continuous business operations through expert agent coordination.
Extended thinking: The strangler fig pattern, named after the tropical fig tree that gradually envelops and replaces its host, represents the gold standard for risk-managed legacy modernization. This workflow implements a systematic approach where new functionality gradually replaces legacy components, allowing both systems to coexist during transition. By orchestrating specialized agents for assessment, testing, security, and implementation, we ensure each migration phase is validated before proceeding, minimizing disruption while maximizing modernization velocity.]
resources/implementation-playbook.md.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 25,173 | 30,272 | +20% | 1 | 1 | 0% | 3,864 | 3,946 | +2% | 0 | 0 | — |
case-02 | fail→fail | 33,329 | 27,928 | -16% | 1 | 1 | 0% | 5,938 | 5,725 | -4% | 0 | 0 | — |
case-03 | fail→fail | 20,715 | 20,762 | +0% | 1 | 1 | 0% | 3,198 | 4,095 | +28% | 0 | 0 | — |
case-04 | pass→pass | 19,226 | 20,065 | +4% | 1 | 1 | 0% | 3,448 | 4,363 | +27% | 0 | 0 | — |
case-05 | pass→pass | 15,573 | 15,787 | +1% | 1 | 1 | 0% | 2,681 | 3,489 | +30% | 0 | 0 | — |
case-06 | pass→pass | 22,053 | 12,734 | -42% | 1 | 1 | 0% | 3,983 | 3,011 | -24% | 0 | 0 | — |
case-07 | fail→fail | 14,017 | 13,251 | -5% | 1 | 1 | 0% | 2,129 | 2,699 | +27% | 0 | 0 | — |
case-08 | fail→fail | 17,250 | 20,655 | +20% | 1 | 1 | 0% | 2,649 | 4,014 | +52% | 0 | 0 | — |
case-09 | fail→fail | 18,341 | 17,238 | -6% | 1 | 1 | 0% | 3,167 | 3,640 | +15% | 0 | 0 | — |
case-10 | fail→fail | 15,325 | 13,828 | -10% | 1 | 1 | 0% | 2,567 | 2,903 | +13% | 0 | 0 | — |
case-11 | fail→fail | 16,251 | 14,506 | -11% | 1 | 1 | 0% | 2,713 | 3,124 | +15% | 0 | 0 | — |
case-12 | fail→fail | 16,704 | 14,093 | -16% | 1 | 1 | 0% | 2,606 | 2,858 | +10% | 0 | 0 | — |
case-13 | fail→fail | 15,422 | 16,229 | +5% | 1 | 1 | 0% | 2,551 | 3,316 | +30% | 0 | 0 | — |
case-14 | fail→fail | 19,821 | 20,161 | +2% | 1 | 1 | 0% | 3,256 | 4,237 | +30% | 0 | 0 | — |
case-15 | fail→fail | 13,438 | 12,671 | -6% | 1 | 1 | 0% | 2,022 | 2,825 | +40% | 0 | 0 | — |
case-16 | fail→fail | 15,424 | 12,485 | -19% | 1 | 1 | 0% | 2,434 | 2,684 | +10% | 0 | 0 | — |
case-17 | fail→fail | 19,497 | 23,085 | +18% | 1 | 1 | 0% | 3,074 | 4,552 | +48% | 0 | 0 | — |
case-18 | fail→pass | 33,363 | 9,878 | -70% | 1 | 1 | 0% | 6,003 | 2,431 | -60% | 0 | 0 | — |
case-19 | fail→fail | 50,529 | 19,632 | -61% | 1 | 1 | 0% | 3,093 | 3,738 | +21% | 0 | 0 | — |
case-20 | fail→fail | 17,187 | 13,933 | -19% | 1 | 1 | 0% | 2,828 | 2,785 | -2% | 0 | 0 | — |
case-21 | fail→fail | 17,102 | 20,249 | +18% | 1 | 1 | 0% | 2,910 | 3,790 | +30% | 0 | 0 | — |
case-22 | fail→fail | 14,031 | 16,032 | +14% | 1 | 1 | 0% | 2,126 | 3,207 | +51% | 0 | 0 | — |
case-23 | fail→fail | 21,301 | 22,196 | +4% | 1 | 1 | 0% | 3,715 | 4,580 | +23% | 0 | 0 | — |
case-24 | fail→fail | 13,444 | 10,993 | -18% | 1 | 1 | 0% | 2,417 | 2,650 | +10% | 0 | 0 | — |
case-25 | fail→fail | 21,734 | 17,836 | -18% | 1 | 1 | 0% | 2,980 | 3,440 | +15% | 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. 25 cases were attempted. The headline lift of +4 percentage points is the difference between those two pass rates over the 25 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.