docs(grok): state the mixed-session drop and the global daily re-derivation plainly

The daily-cache re-derivation test seeded v18, a version that only ever
existed as an unreleased draft of this change. Seed the shipped v17 so the
test models the 17 -> 19 upgrade path users actually hit, and rename it: the
bump re-derives every day for every provider, not just Grok, because the
daily cache has no per-provider invalidation. The Grok day stays as the
fixture since Grok is what the bump exists to correct.

The changelog entry now says outright that Grok totals change materially on
upgrade (150K -> 96.3M cache-read tokens on a 568-session corpus), that a
turn without a turn_completed record inside an otherwise-covered session is
dropped rather than estimated, and that the one-time daily re-derivation
reads the warm session cache and keeps the superseded file. The
context-bloat denominator fix moves to Fixed and names the providers it
corrects.

docs/providers/grok.md gets the same undercount warning in the token model
and a matching entry under Quirks.

docs/providers/grok.md

@@ -21,7 +21,7 @@ JSON + JSONL. `summary.json` holds the session id, cwd, timestamps, and `current
 
 **Estimated fallback.** Older sessions without any valid `turn_completed.usage` record use the running context fill (`totalTokens` per chunk) and the existing compaction-aware per-turn curve. That path remains flagged `costIsEstimated`; a completed record is never blended with the heuristic.
 
-If a session spans a CLI upgrade and has both completed records and streamed turns without a matching record, the all-or-nothing authoritative path keeps the recorded totals, drops the uncovered turns, and marks the emitted row `costIsEstimated: true`. A later parse can fill an open turn once it writes a record; pre-upgrade turns that never do are dropped.
+**Mixed sessions undercount.** The choice between the two paths is per session, not per turn. If a session has at least one usable `turn_completed` record, the whole session is billed from the summed records and any turn WITHOUT a record contributes nothing at all - its tokens are dropped, not estimated, so such a session reads low. The row is marked `costIsEstimated: true` rather than claiming full provider coverage. This is deliberate: blending the heuristic into real records would reintroduce the roughly 5x output over-count this parser exists to remove. It happens when a session straddles a CLI upgrade or a run dies before writing its last record; an open turn is filled by a later parse once it writes one, pre-upgrade turns never are. Measured on a 568-session corpus, 1 turn out of 566 was uncovered.
 
 ## Pricing
 
@@ -38,6 +38,7 @@ Per `grok:<session-dir>:<updated_at>:<id>`.
 ## Quirks
 
 - **Two token paths.** Completed turns carry provider usage; sessions from older CLI versions have only the context curve and therefore remain estimates (likely an upper bound, since re-sent context is cached server-side and not exposed in those files).
+- **A turn with no `turn_completed` record is dropped inside an otherwise-covered session** (see Token model). The session still reports, marked estimated, but reads low by those turns.
 - **No bash-command capture.** Tool names come from `signals.toolsUsed`; per-command bash text is not extracted, so `bashCommands` is empty.
 - **Whole-session timestamp.** Spend is attributed to `updated_at`, since the context curve is cumulative.
 - **Subscription vs API.** Grok Build runs via either a metered xAI API account (tiered) or a SuperGrok subscription; the session files do not record which.