Code Cleanup

To remove deprecated shims and dead exports left by AI agents, scan scenario files for legacy markers, perform exhaustive usage searches, run test suites, and document verification before deleting the code.

Finding stale TODO comments and legacy markers for a code refactor involves detecting tags like deprecated, legacy, Old, and V1 across scenario-specific files while excluding shared packages to identify safe removal candidates.

To safely prune dead code without breaking cross-scenario imports, execute cross-scenario import checks, validate usages, and run test suites prior to eliminating any candidate code to prevent regressions.

Yes, you can clean up forwarding shims and compatibility re-exports in scenario files without affecting shared packages by targeting analysis exclusively at scenario-specific files and excluding shared code from deletion.

Track code cleanup progress and prioritize high-impact files for dead code removal by logging visits, excluding already clean files, and documenting residual questions for future review during the refactor process.

You should not use automated code cleanup for removing deprecated implementations when exhaustive usage searches and test runs cannot be performed, as deleting code without verified removal workflows risks regression.