The on-device index was fetched once and never revalidated: the only
freshness check was two existsSync calls. A move added after that fetch was
invisible to meaning search permanently, not down-ranked but absent from the
candidate set. The registry manifest on the same command carries a 24h TTL,
so the two halves of one feature disagreed about staleness.
The dropped count reported over-coverage only, names the index has that the
registry lacks. Under-coverage was never computed, so the harmless direction
was instrumented and the costly one was silent. Reproduced with an index
truncated to 120 of 168 moves: dropped read 0, perfect health, while 48
moves were unreachable.
Counts under-coverage from the name list the artifact already carries, so no
extra request. Warns only when non-zero, and names the remedy.
The remedy had to be made true: --on-device could not refresh a stale index
because hasLocalVectors short-circuited the fetch. That flag now refetches
when the index is absent or no longer covering.
Two defects the reproduction surfaced. A failed refresh reported the tier
unavailable while the old vectors were still on disk and still ranking. And
the fetch wrote its two files one at a time, so failing between them paired
a new name list with an old matrix, a hard load error rather than stale
data. It now writes both or neither, which matters more once refresh runs on
staleness.
top_score returns, scoped to the on-device tier and set to the score of the
best result actually shown rather than the ranking head, which can describe
a row the caller never received.
Also: scripts/ is now typechecked. It never was, which is how a build script
that crashes after the paid embedding call, and two scripts whose imports do
not resolve at all, went unnoticed. 43 errors fixed, no suppressions.
And the docs stop describing a --smart hosted tier that was deleted, an
item that does not exist, and a registry refresh that cannot fix a stale
vector index.
## What
- Add `bun run release:prepare <version>` as the maintainer-facing stable release entrypoint.
- Make the first run draft missing changelog artifacts and intentionally exit before tagging; rerunning after manual review delegates to `set-version`.
- Tighten the direct `set-version` guard so stable releases also fail when generated TODO changelog copy is still present.
- Update maintainer docs to recommend `release:prepare` while keeping `changelog:draft` as the lower-level regeneration tool.
## Why
Stable releases should be hard to run without reviewed GitHub release notes and Mintlify changelog copy. This keeps the existing manual rewrite step, but makes the expected path one command that engineers can rerun after review.
## How
- Added `scripts/release-prepare.ts` with parsing, draft/review/set-version action selection, and command forwarding.
- Added focused script tests for parser behavior, action selection, command forwarding, and TODO detection.
- Extracted shared script CLI parsing helpers so `changelog:draft` and `release:prepare` use the same option handling.
- Adjusted `changelog:draft --write` so an existing release file is left unchanged unless `--force` is passed, while still allowing a missing docs entry to be added.
## Test plan
- [x] Unit tests added/updated: `bun run test:scripts`
- [x] Format check: `bun run format:check`
- [x] Lint: `bun run lint`
- [x] Typecheck: `bun run --filter '*' typecheck`
- [x] Fallow audit: `bunx fallow audit --base origin/main --fail-on-issues`
- [x] Manual CLI checks: `bun run release:prepare --help`; `bun run set-version 9.9.9` fails before mutation when changelog artifacts are missing
- [x] Documentation updated
Miguel Ángel