Refactor code safely with mandatory baseline tests and Gradle verification.
npx skills add https://github.com/abnegate/claudes --skill refactor-abnegate
Please help me install this Agent Skill. Skill: refactor Source: https://github.com/abnegate/claudes/tree/main/skills/refactor Command: npx skills add https://github.com/abnegate/claudes --skill refactor-abnegate
It prevents risky refactors by enforcing that all tests pass both before and after changes, reducing the chance of subtle behavior regressions.
Ask the AI to refactor the targeted code described as file(s) and intent, for example: "Refactor $ARGUMENTS while ensuring no behavior changes and that all Gradle tests pass before and after."
High-intent search queries and answers about installing and using this skill.
To safely refactor code without breaking behavior, run the full test suite before and after modifications to establish a baseline. Identifying public APIs and writing characterization tests for uncovered branches locks in current functionality during changes.
Characterization tests lock in the current behavior of existing code before you refactor it. You need them when test coverage is insufficient, ensuring untested edge cases and error paths are documented to prevent regressions during structural changes.
To ensure Gradle test coverage is sufficient before a module refactor, generate a baseline coverage report and run the full test suite. This identifies uncovered branches and missing edge-case tests that must be addressed before making safe changes.
Yes, you can execute cross-cutting refactor requests using isolated parallel worktrees. This planner and verifier workflow validates the refactor plan in separate architect worktrees, allowing safe stepwise changes before consolidating the final diff.
The best way to manage safe code changes with untested error paths is to first fill the test gaps with prioritized characterization tests. This discovers and locks in the current behavior of error paths before executing any stepwise refactoring workflow.
Your refactoring workflow fails because koverReport coverage thresholds enforce mandatory baseline testing standards. This prevents risky refactors by ensuring all public APIs and edge cases are adequately tested before and after code modifications.
AtyantikAnalyze coverage reports to identify uncovered lines and missing branches.
benediktmsRefactor code with minimal mechanical changes and test suite validation.
TheBushidoCollectiveAdd characterization tests before refactoring untested legacy code.
whmathews15Create and manage isolated Git worktrees under .worktrees.
