Record Aug 15 measurement showing Phases 1–2 missed success targets, Phase 3 code is live with 15% embedding backfill and zero L0 traffic yet. Co-authored-by: Cursor <cursoragent@cursor.com>
* docs(convoy): seed scanner identify upgrade epic and sub-convoys Baseline scan_attempts telemetry and three-phase plan for faster, more accurate card identification without touching scanner chrome. Co-authored-by: Cursor <cursoragent@cursor.com> * feat(scanner): tighten Layer-1 identify hot path (Phase 1) Cut verify hold-still gates, OCR collector numbers on Layer 1, request structured Gemini JSON, and skip automatic L2 refine when L1 opens the printing picker. Includes convoy UX/architecture briefs and unit tests. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>