Archive index access (opt-in)

Archive index access (opt-in)

The knowledge-hub-archive sibling repo holds Class-4 cold-storage artefacts removed from the public knowledge-hub repo during the ID-68 doc-lifecycle relocation (audits, testing/uat, point-in-time operations, archived research; see the archive repo’s PROVENANCE-ID-68.28.md).

By design (TECH PC-5 index-isolation, AC-A4) the code-intelligence tools — GitNexus and ccc — do NOT index the archive corpus. Searches run from the main knowledge-hub repo return zero archive-path results. This keeps stale point-in-time content out of live impact analysis and concept search.

When you need the archive in scope

Archive content is point-in-time and possibly stale — the public docs/reference/*.json ledgers remain the source of truth for current status (TARGET-STATE §3.4). Only opt the archive into an index when you are explicitly researching a past decision and accept the staleness.

GitNexus

Point GitNexus at the archive checkout on demand and analyse it as a separate repo:

# from the archive checkout
cd /path/to/knowledge-hub-archive
bun run gitnexus:analyze   # or the gitnexus CLI 'analyze' against this path

Treat the resulting index as a throwaway, point-in-time snapshot. Do not wire it into the main repo’s standing index or CI.

ccc

Run a ccc index/search scoped to the archive checkout path on demand, then discard it. Do not add the archive path to the main repo’s standing ccc index.

Invariant

The opt-in nature is the invariant: archive content must never silently enter the main repo’s standing GitNexus/ccc index. If an impact-analysis or concept-search result ever surfaces an audits/, testing/uat/, point-in-time operations/, or archived research/ path from the main repo, that is an index-isolation regression — re-clean the main index.