Skip to content

Commit 22422be

Browse files
author
jun
committed
docs(cursor): record the measured -fast round-trip that proves audit B11
The reviewer's claim that a bare -fast suffix picks the wrong dimension was not theoretical. Measured: claude-opus-5-fast resolves to claude-opus-5-high-fast (clamped, and in the quarantined regular family) while the Codex toggle would send claude-opus-5-thinking-max-fast. The mirror case is just as wrong -grok-4.6-thinking-fast degrades to a bare grok-4.6 with no effort and no fast marker, because grok has no thinkingFast spec. Either fixed suffix is wrong for half the table, which is why cursorFastIdFor composes from the base's defaultVariant.
1 parent 7adb1e6 commit 22422be

1 file changed

Lines changed: 30 additions & 0 deletions

File tree

‎devlog/_plan/260902_cursor_unified_identity/002_audit_round1.md‎

Lines changed: 30 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -226,3 +226,33 @@ that call site are one atomic edit, and the Grok accept-row test belongs to WP3.
226226
in `040_residuals.md` when WP4 lands rather than leaving it only in prose.
227227

228228
Reviewer's normalized line: `VERDICT: GO-WITH-FIXES (blockers=10)`.
229+
230+
## B11 confirmed by measurement (`.tmp/probe3.ts`, at 7adb1e66a)
231+
232+
The reviewer's claim was not theoretical. Bare `-fast` on a thinking-default base picks the
233+
REGULAR-fast sibling and diverges from what the Codex toggle would send:
234+
235+
```
236+
base default listed id kind resolved wire (max)
237+
claude-opus-4-7 thinking claude-opus-4-7-fast fast claude-opus-4-7-max-fast
238+
claude-opus-4-7-thinking-fast thinkingFast claude-opus-4-7-thinking-max-fast
239+
umbrella claude-opus-4-7 thinking claude-opus-4-7-thinking-max
240+
claude-opus-5 thinking claude-opus-5-fast fast claude-opus-5-high-fast <- clamped AND quarantined family
241+
claude-opus-5-thinking-fast thinkingFast claude-opus-5-thinking-max-fast
242+
grok-4.5 regular grok-4.5-fast fast grok-4.5-high-fast
243+
grok-4.5-thinking-fast thinkingFast grok-4.5 <- degrades to a bare id
244+
grok-4.6 regular grok-4.6-fast fast grok-4.6-xhigh-fast
245+
grok-4.6-thinking-fast thinkingFast grok-4.6 <- degrades to a bare id
246+
```
247+
248+
Two consequences the fix must respect, both visible above:
249+
250+
1. For a thinking-default base, only `<base>-thinking-fast` round-trips to the variant the
251+
toggle picks. `claude-opus-5-fast` additionally clamps max->high and lands in the
252+
quarantined regular family.
253+
2. For a regular-default base, `<base>-thinking-fast` is WRONG the other way: grok has no
254+
thinkingFast spec, so `resolveCursorSelection` falls back to `variants.regular` and emits
255+
a bare `grok-4.6` with no effort and no fast marker at all.
256+
257+
So the id must be composed per base from `defaultVariant`, exactly as `cursorFastIdFor` in
258+
030 §1 now does — a single shared suffix would be wrong for one half of the table either way.

0 commit comments

Comments
 (0)