Full Codebase Migrator Skill for Claude

Full Codebase Migrator

Leverage the full 1M token context window to ingest an entire codebase, understand its architecture end-to-end, and produce a comprehensive, file-by-file migration plan that a team can execute sequentially without conflicts.

Contents

• references/migration-types.md -- supported migration types; when to use and when not to use this skill.
• references/ingestion.md -- architecture, glob patterns, metadata collection, read order, context budgeting, and dependency graph construction (Steps 0-2).
• references/workflow-detail.md -- per-file assessment, migration order, risk scoring, and effort estimation templates (Steps 3-6).
• references/output-template.md -- migration-plan.md and migration-plan.json templates, edge cases, and the final quality checklist (Steps 7-8).

Workflow

• Identify scope. Determine migration type, scope (full repo or directory), output location for migration-plan.md, and any exclusions or constraints. Infer from the user's prompt when already specified. See references/migration-types.md for supported types and fit.
• Ingest the codebase. Glob all source files, collect metadata with Bash (line counts, package manifest, configs, git history, directory tree), then Read every source file. For 500+ file codebases, dispatch Agent sub-agents to read directory subtrees in parallel. See references/ingestion.md.
• Build the dependency graph. Extract every import, build an adjacency list, classify each file into layers, detect circular dependencies, and classify external dependencies. See references/ingestion.md.
• Assess each file. Produce a per-file migration assessment covering layer, complexity, patterns found, required changes, prerequisite dependencies, risk factors, and testing impact. See references/workflow-detail.md.
• Calculate migration order. Run a topological sort by layer, apply practical adjustments (quick wins, high-risk-early, break cycles first), and group files into buildable, PR-sized phases. See references/workflow-detail.md.
• Assess risk. Score each file 1-5 on complexity, centrality, volatility, test coverage, and external coupling. Build the migration-level risk matrix and rollback strategy. See references/workflow-detail.md.
• Estimate effort. Derive per-file, per-phase, and total estimates with overhead and a 20% buffer, then translate into calendar time by team size. See references/workflow-detail.md and the effort calibration table in references/migration-types.md.
• Generate the deliverable. Write migration-plan.md to the output location, and migration-plan.json when machine-readability helps. Verify against the quality checklist before delivering. See references/output-template.md.

Edge Cases

Codebases too large for context, monorepos, partially migrated codebases, and codebases without tests each need adjusted handling. See references/output-template.md.