Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Migration workflow - research → analyze → plan → implement → review
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-11 | ✗→✓ | ▲ Improved | 146% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 59% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 83% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 57% | 0% |
Safe migrations for frameworks, languages, and infrastructure.
┌──────────┐ ┌──────────┐ ┌────────────┐ ┌──────────┐ ┌───────────┐
│ oracle │───▶│ phoenix │───▶│ plan- │───▶│ kraken │───▶│ surveyor │
│ │ │ │ │ agent │ │ │ │ │
└──────────┘ └──────────┘ └────────────┘ └──────────┘ └───────────┘
Research Analyze Plan Implement Review
target current migration changes migration| # | Agent | Role | Output | |---|-------|------|--------| | 1 | oracle | Research target framework/version | Research report | | 2 | phoenix | Analyze current codebase for migration impact | Impact analysis | | 3 | plan-agent | Create phased migration plan | Migration plan | | 4 | kraken | Implement migration changes | Code changes | | 5 | surveyor | Review migration for completeness | Migration review |
Migrations are high-risk:
The extra research and review phases catch issues early.
Task(
subagent_type="oracle",
prompt="""
Research migration target: [TARGET]
Investigate:
- Breaking changes from current version
- New APIs and patterns
- Deprecated features we use
- Migration guides from official docs
- Common pitfalls and solutions
Output: Migration research report
"""
)Task(
subagent_type="phoenix",
prompt="""
Analyze codebase for migration: [FROM] → [TO]
Identify:
- Files using deprecated APIs
- Dependency conflicts
- Patterns that need updating
- Test coverage of affected areas
- Risk areas (critical paths)
Output: Impact analysis with affected files
"""
)Task(
subagent_type="plan-agent",
prompt="""
Create migration plan: [FROM] → [TO]
Research: [from oracle]
Impact: [from phoenix]
Plan should:
- Be phased (incremental if possible)
- Each phase independently testable
- Include rollback strategy
- Prioritize critical path stability
Output: Phased migration plan
"""
)Task(
subagent_type="kraken",
prompt="""
Implement migration phase: [PHASE_N]
Plan: [from plan-agent]
Requirements:
- Follow plan exactly
- Run tests after each change
- Document any deviations
- Stop if tests fail
Output: Completed phase with test results
"""
)Task(
subagent_type="surveyor",
prompt="""
Review migration: [FROM] → [TO]
Check:
- All deprecated APIs replaced
- No remaining compatibility shims
- Tests passing
- Performance acceptable
- No security regressions
Output: Migration review report
"""
)User: /migrate React 17 → 18
→ Full workflow with hooks/concurrent mode changesUser: /migrate Python 3.9 → 3.12
→ Focus on new syntax, removed featuresUser: /migrate MongoDB → PostgreSQL
→ Schema design, data transformation, query rewriteUser: /migrate AWS → GCP
→ Service mapping, config translationUser: /migrate from Express to Fastify
Claude: Starting /migrate workflow: Express → Fastify
Phase 1: Researching Fastify...
[Spawns oracle]
Key differences:
- Schema-based validation (vs middleware)
- Plugin architecture
- Async/await native
- Different middleware pattern
Phase 2: Analyzing current Express usage...
[Spawns phoenix]
Impact analysis:
- 23 route files to update
- 8 middleware to convert
- 3 plugins need Fastify equivalents
- Test coverage: 78%
Phase 3: Creating migration plan...
[Spawns plan-agent]
Migration Plan: Express → Fastify
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Phase 1: Setup & Core
- Add Fastify alongside Express
- Migrate core server setup
- Run both in parallel
Phase 2: Routes (by domain)
- Migrate auth routes
- Migrate user routes
- Migrate product routes
- ...
Phase 3: Middleware
- Convert error handler
- Convert auth middleware
- Convert logging
Phase 4: Cleanup
- Remove Express
- Remove compatibility layer
- Final testing
Rollback: Keep Express until Phase 4 complete
Phase 4: Implementing Phase 1...
[Spawns kraken]
✅ Fastify server running alongside Express
✅ Core routes migrated
✅ Tests passing
[Continues through phases...]
Phase 5: Reviewing migration...
[Spawns surveyor]
✅ All Express code removed
✅ 23/23 routes migrated
✅ Performance improved (2.1x faster)
✅ No security regressions
Migration complete! Express → FastifyFor large codebases, run one phase at a time:
User: /migrate React 17 → 18 --phase 1
[Runs only Phase 1]
User: /migrate React 17 → 18 --phase 2
[Runs Phase 2, reads previous handoff]--phase N: Run specific phase only--dry-run: Plan without implementing--rollback: Execute rollback plan--parallel: Run new alongside old (strangler fig)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-13 | fail→fail | 11,093 | 6,964 | -37% | 1 | 1 | 0% | 1,812 | 2,636 | +45% | 0 | 0 | — |
case-11 | fail→pass | 5,680 | 4,793 | -16% | 1 | 1 | 0% | 956 | 2,350 | +146% | 0 | 0 | — |
case-01 | fail→pass | 23,757 | 7,957 | -67% | 1 | 1 | 0% | 4,729 | 2,533 | -46% | 0 | 0 | — |
case-02 | fail→fail | 28,593 | 6,912 | -76% | 1 | 1 | 0% | 5,950 | 1,978 | -67% | 0 | 0 | — |
case-03 | fail→fail | 28,415 | 7,014 | -75% | 1 | 1 | 0% | 6,199 | 2,488 | -60% | 0 | 0 | — |
case-04 | fail→pass | 6,282 | 2,726 | -57% | 1 | 1 | 0% | 1,176 | 1,875 | +59% | 0 | 0 | — |
case-05 | fail→pass | 6,172 | 2,791 | -55% | 1 | 1 | 0% | 1,087 | 1,988 | +83% | 0 | 0 | — |
case-12 | fail→pass | 7,522 | 2,646 | -65% | 1 | 1 | 0% | 1,224 | 1,925 | +57% | 0 | 0 | — |
case-06 | fail→fail | 11,245 | 2,620 | -77% | 1 | 1 | 0% | 1,965 | 1,716 | -13% | 0 | 0 | — |
case-07 | fail→pass | 12,427 | 1,535 | -88% | 1 | 1 | 0% | 2,201 | 1,704 | -23% | 0 | 0 | — |
case-08 | fail→pass | 7,614 | 2,862 | -62% | 1 | 1 | 0% | 1,282 | 1,978 | +54% | 0 | 0 | — |
case-09 | fail→pass | 10,856 | 2,816 | -74% | 1 | 1 | 0% | 1,856 | 1,954 | +5% | 0 | 0 | — |
case-10 | fail→pass | 6,546 | 2,951 | -55% | 1 | 1 | 0% | 1,013 | 1,905 | +88% | 0 | 0 | — |
case-14 | fail→pass | 17,617 | 12,262 | -30% | 1 | 1 | 0% | 3,219 | 3,480 | +8% | 0 | 0 | — |
case-15 | fail→fail | 4,420 | 3,714 | -16% | 1 | 1 | 0% | 654 | 2,034 | +211% | 0 | 0 | — |
case-16 | fail→fail | 10,894 | 5,629 | -48% | 1 | 1 | 0% | 1,981 | 2,480 | +25% | 0 | 0 | — |
case-17 | pass→fail | 21,046 | 7,313 | -65% | 1 | 1 | 0% | 4,018 | 2,373 | -41% | 0 | 0 | — |
case-18 | fail→pass | 20,082 | 17,175 | -14% | 1 | 1 | 0% | 3,158 | 4,774 | +51% | 0 | 0 | — |
case-19 | pass→pass | 8,402 | 9,382 | +12% | 1 | 1 | 0% | 1,355 | 3,069 | +126% | 0 | 0 | — |
case-20 | pass→pass | 9,935 | 14,345 | +44% | 1 | 1 | 0% | 2,014 | 1,825 | -9% | 0 | 0 | — |
case-21 | fail→fail | 3,354 | 1,770 | -47% | 1 | 1 | 0% | 587 | 1,669 | +184% | 0 | 0 | — |
case-22 | fail→pass | 10,699 | 4,075 | -62% | 1 | 1 | 0% | 1,787 | 2,231 | +25% | 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 +50 percentage points is the difference between those two pass rates over the 21 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/29/2026 | +55% |
Other measured skills in the registry, with their headline benchmark lift.