Skip to content

v1.3.0 — abstract index + fast retrieval (cz_get)

Choose a tag to compare

View all tags
@collincusce collincusce released this 28 Jun 07:49
· 227 commits to main since this release
v1.3.0
4be4916

Fast retrieval — a deterministic abstract index over the memory corpus, so an agent reads exactly the entry it needs instead of loading whole files.

  • cz_get(id) — addressable single-entry fetch (D-/INVARIANT-/H-/L- ids); the full body is resolved from canonical markdown on demand. Read-only.
  • Abstracts on cz_analyze — each ranked hit carries a one-line abstract; a pre-registered cost experiment measured a 48.3% mean payload-token reduction per lookup at equal answer accuracy (deterministic, no live LLM).
  • Write-time near-duplicate-lesson advisory on cz_add_lesson (length-normalized Jaccard): nudges consolidation instead of appending. Advisory only, never blocks; the corpus stays append-only.
  • Upgrade path — clauderize init and clauderize reindex build/refresh the gitignored abstract index idempotently; clauderize doctor detects a missing or schema-stale cache and advises reindex.

Markdown stays canonical (INVARIANT-01); the index is a disposable rebuilt-from-markdown cache, with no new runtime dependency. Tool surface 41 to 42 (cz_get).

Assets 2
Loading