Tip: Use format_refactor_plan_report() to read results outside the conversation

Recommendation

When you need the refactor-plan output in a script or CI pipeline, call format_refactor_plan_report(result, input_data) from workflows.refactor_plan_report directly. It converts the raw WorkflowResult dict into a human-readable report with the same Summary, Refactoring, and Suggestions sections you see in a Claude Code conversation.

Why: The three internal subagents (debt-scanner, impact-analyzer, plan-generator) are underscore-prefixed internals that can change without notice. format_refactor_plan_report() is the stable surface for consuming their synthesized output.

Tradeoff: The function expects the result dict that RefactorPlanWorkflow.execute() produces. If you pass a partial or hand-constructed dict, sections may render incomplete. Run execute() first; don't shortcut by constructing the dict yourself.

Source files

Tags: refactor, tech-debt, complexity