Modernize Without the Meltdown

Legacy code is the backbone of business logic, but refactoring it often feels like defusing a bomb in the dark. This skill provides a disciplined, low-risk methodology for modernizing legacy codebases using a "safety-first" developer mindset. It prevents the common pitfalls of AI-driven rewrites by enforcing a strict, multi-phase process that prioritizes stability over speed.

What it does

The skill guides your AI agent through a rigorous modernization workflow:

• Characterization Mapping: Identifies boundaries, side effects, and "sacred" business logic before a single line is changed.
• Safety Net Construction: Automatically generates characterization tests that capture current behavior (even the "bugs" that users rely on).
• Incremental Migration: Implements proven patterns like Strangler Fig and Branch by Abstraction to phase out old code safely.
• Gatekept Execution: Prevents the deletion of old code until new paths are verified and the user gives explicit sign-off.

Why developer-led agents need this

Standard AI prompting often leads to "Big Bang" rewrites that lose subtle business rules. This skill transforms your agent into a careful surgeon, ensuring every change has a rollback path and every refactor is backed by objective proof. It is compatible with any framework and excels in Grok-driven environments where deep context is key.