refactor-archive

Archiving a refactor project involves moving it under refactor/archive and updating all self-references in configuration files like refactor.toml and program.md to point to the new location. Durable artifacts such as results.tsv and last_eval.json are preserved while transient logs are pruned.

Maintaining a curated artifact bundle after refactor closeout requires parking the finished loop under refactor/archive and updating the corresponding archive note. This preserves durable runtime artifacts like last_eval.json while pruning transient logs for a clean historical record.

Yes, archiving a refactor project rewrites self-references in configuration files like refactor.toml and program.md. This ensures all internal paths reflect the archived location under refactor/archive, maintaining consistency across the version control workflow.

You need to move finished refactors into a durable archive when you require a complete historical record after project closeout, when parking a finished development loop, or when maintaining curated artifact bundles across multiple version control branches.

When relocating a refactor to the archive directory, durable artifacts like results.tsv and last_eval.json are preserved for long-term traceability, while transient runtime logs are pruned to reduce clutter and maintain a clean archival record.

Updating archive notes under .claude/agent-memory/refactor-corpus/archive/ improves retrieval by maintaining a centralized, curated record of archived refactors. This ensures long-term accessibility and traceability of the historical project data across branches.