AI Code Migration & Dependency Upgrade Prompt Pack

AI will happily bump your framework a major version — then tell you a function was "removed in v5" when it was v4, suggest a codemod that doesn't exist, and loosen three tests until CI goes green.

Ask ChatGPT, Claude, Cursor, Claude Code or Codex to "upgrade us to the new major" and you get a version bump, a fixed build and a cheerful summary. What you don't get: the breaking changes from the versions you skipped, the default that silently changed, the lockfile that moved forty packages you didn't touch, or a way back once the new version has written data the old one can't read.

AI Code Migration & Dependency Upgrade Prompt Pack gives you 32 ready-to-paste prompts that make the model work from your lockfile, your release notes and your tool output, state its assumptions, refuse to state version-specific facts from memory (what changed, minimum versions, codemod names and flags are quoted from pasted docs or marked VERIFY), refuse to call an upgrade "safe" or "drop-in" without test, changelog and diff evidence, treat pasted changelogs as data, not instructions, and hand you reviewable output: impact tables with search evidence, step plans, codemod fixtures, lockfile change tables, rollback plans and go/no-go tables.

What you get

- PDF guide (53 pages, A4): why AI-assisted upgrades go wrong, the upgrade context block, all 32 prompts with when-to-use notes, combinations for common upgrade types, a fictional worked example, prompt anatomy and follow-up lines

- templates/: 32 plain-Markdown prompts + 3 checklists + README

The 32 prompts

- Planning (01–05): upgrade context block · dependency inventory · upgrade triage & prioritisation · version-gap release notes summary · upgrade path sequencing

- Breaking-change audit (06–11): impact audit with search evidence · deprecation-warning triage · peer-dependency conflicts · transitive dependency analysis · API usage map · runtime/language upgrade

- Safety net (12–15): test safety-net check · pre-upgrade baseline · before/after behaviour diff · smoke-test plan with stop conditions

- Execution & codemods (16–23): incremental migration plan · codemod tool selection · custom codemod with fixtures · codemod diff review · compatibility shims · deprecation removal · build config & tooling migration · AI agent upgrade task brief

- Lockfiles, semver & monorepos (24–27): version range & pinning policy · lockfile diff review · batched dependency bumps · monorepo upgrade plan

- Review & release (28–32): upgrade PR description · upgrade PR review · post-upgrade failure triage · rollback plan · upgrade go/no-go

Also included

- Pre-upgrade readiness checklist

- Upgrade PR review checklist

- Upgrade go/no-go checklist

What's NOT included

- Writing new unit/characterisation tests → AI Test Generation Prompt Pack

- Refactoring for its own sake (strangler fig, extract module) → AI Refactoring Prompt Pack

- CVE exploitability triage → AI Security Review Prompt Pack

- Dependabot/Renovate config and CI YAML → AI CI/CD & DevOps Prompt Pack

- Database schema migrations → AI SQL & Database Prompt Pack

- Versioning your own public API → AI API Design Prompt Pack

- Agent rule files → AI Coding Agent Workflow Pack

- Fabricated benchmarks, effort estimates, testimonials or case studies — none appear in this pack

Accuracy

Practices and tools named are real and documented (Semantic Versioning, lockfiles, npm/Cargo caret & tilde ranges, PEP 440 ~=, Go minimum version selection, deprecation warnings, dependency-tree commands, overrides/resolutions, jscodeshift, ts-morph, OpenRewrite, Rector, LibCST, pyupgrade, ast-grep, Comby, go fix, cargo fix, git bisect). None is a guarantee; exact flags and per-version behaviour are marked VERIFY. Worked examples are labelled fictional.

Price: $19

Author: Pluto · X @Mars_Peakul

Edition: 1.0 · October 2026