team-migration_skill

Overview

This skill automates creation and orchestration of a parallel migration team tailored to a detected technology transition. It analyzes the migration scope, selects specialists from a selection matrix, spawns them in parallel, and leads synthesis into a phased migration plan with verification and rollback procedures. The goal is safe, incremental migrations with clear responsibilities and automated transformations where possible.

How this skill works

The skill parses provided arguments or repository context to detect the migration scope (PR, issue, commit, diff, branch, files, or free-text). It classifies the migration type (language/runtime, framework/library, API protocol, database, build/tooling, infrastructure, or mixed) and applies the specialist selection matrix to decide which experts to spawn. It then issues a single parallel Task call to create the selected subagents, collects their outputs, and synthesizes a migration status report that includes phased plans, verification results, manual tasks, and rollback instructions.

When to use it

• Upgrading runtime, language, or compiler versions across a repository
• Replacing or substantially upgrading frameworks or libraries
• Changing API protocols (REST ↔ GraphQL, HTTP ↔ gRPC)
• Migrating database engines, ORMs, or performing large schema changes
• Switching build systems or major tooling changes (npm → pnpm, Webpack → Vite)
• Replatforming infrastructure or splitting a monolith into services

Best practices

• Always provide the full target context (diffs, files, or descriptive text) so specialists have accurate inputs
• Prefer including conditional specialists when signals are ambiguous to reduce migration risk
• Run all subagent Task calls in one message to ensure parallel execution and consistent context
• Treat the lead as the single synthesizer: subagents return findings but do not synthesize final plans
• Define clear verification criteria and rollback triggers before applying automated transformations

Example use cases

• Automate a Python 2 → 3 migration: detect language signals, spawn code transformer, test migrator, data migrator, and rollback planner
• Migrate REST APIs to GraphQL: select protocol-focused specialists and compatibility bridge builder to enable incremental rollout
• Upgrade a Java monorepo from Java 8 → 17: enlist breaking change analyst, dependency resolver, code transformer, and test migrator
• Move from MySQL to PostgreSQL: spawn data migrator, compatibility bridge builder, and rollback planner to preserve integrity
• Replace Webpack with Vite in a front-end repo: include build/tooling specialists, code transformer for config changes, and test migrator