Research Archive Skill

Archive a completed research plan after findings have been synthesized and consumed.

Process

1. Identify the research to archive:

- If the user specifies a number/name, use that - Otherwise, scan .gumbo/research/NNNN-*/ (exclude archive/ and unnumbered legacy directories) for research with status "synthesized" in .research-state.json - If multiple candidates, ask user to choose

1. Verify completion:

- Check that synthesis.md exists - Check that findings files exist for each question - If synthesis is missing, ask: "Research has no synthesis yet. Archive anyway?"

1. Update research files:

Update research-plan.md status: ```markdown ## Status: ARCHIVED

Archived: YYYY-MM-DD ```

Update .research-state.json: ``json { "status": "archived", "archivedat": "2026-01-28T10:30:00Z", "updatedat": "2026-01-28T10:30:00Z", ...existing fields... } ``

1. Move to archive:

``bash mv .gumbo/research/NNNN-topic-name .gumbo/research/archive/ ``

1. Confirm to user:

`` Archived: .gumbo/research/archive/NNNN-topic-name/ Status: ARCHIVED Questions: N answered Synthesis: Yes/No Archived: YYYY-MM-DD ``

Notes

• Research referenced by implementation plans (via relative paths like ../../.gumbo/research/NNNN-topic/) will have broken links after archiving. The archive path becomes ../../.gumbo/research/archive/NNNN-topic/. Consider this before archiving research that active plans reference.
• Unnumbered legacy research directories (e.g., dagre-layout/, edge-routing-deep-dive/) should not be archived through this skill — they predate the numbering convention.