diff --git a/analysis/2026-10-07_report-design-review.md b/analysis/2026-10-07_report-design-review.md
index c15c1b1..99bb3f6 100644
--- a/analysis/2026-10-07_report-design-review.md
+++ b/analysis/2026-10-07_report-design-review.md
@@ -1,5 +1,5 @@
 # Per-genome HTML/PDF report (`matpredict report genome`): build and design review
-Status: open (implemented on branch `report-genome`; awaiting curator review)
+Status: decided (PR #48; design questions 1-3 settled 2026-10-07)
 
 ## Question
 Can one `detect` run be turned into a report that a biologist reads in five seconds and a curator can audit, that
@@ -81,7 +81,8 @@ What the report now does:
 ## Curator decisions
 Decided 2026-10-07: (1) `detect` writes report.html by default, opt out with `--no-html` / `MATPREDICT_HTML=0`
 (set in the batch scripts). (2) PDF engine: WeasyPrint in the pixi environment and image (25 packages, 10.3 MB
-download, against about 300 MB for chrome-headless-shell 154 and its libraries); revisit if insufficient. Open: (3) Withheld loci collapsed (current) or hidden by default?
+download, against about 300 MB for chrome-headless-shell 154 and its libraries); revisit if insufficient. (3) Withheld loci: kept as now (curator 2026-10-07): an always-visible section with count, explanation and
+reasons; the table collapsed on screen and expanded in print/PDF.
 
 ## Files
 - Code: `src/MATPredict/report/`, `src/MATPredict/detect/provenance.py`; tests `tests/report/`,
diff --git a/analysis/INDEX.md b/analysis/INDEX.md
index c5a20ea..d103ea5 100644
--- a/analysis/INDEX.md
+++ b/analysis/INDEX.md
@@ -71,6 +71,10 @@ in yet; superseded = replaced by a later study. `results/` = main checkout.
 ## Reads-only typing
 | Date | Report | Status | Key numbers |
 |---|---|---|---|
-| 10-07 | [Per-genome HTML/PDF report: design review](2026-10-07_report-design-review.md) | open | 3 review rounds; round 1 6 P0, round 2 1 P0; Chrome and WeasyPrint PDFs; 30 tests |
 | 10-06 | [Fola reads-type](2026-10-06_fola-reads-type.md) | open | 147/148 agree with samtools breadth using a Fola-derived panel (145/148 with GenBank); 50a locus assembled (MAT1-2) |
 | 10-07 | [A. fumigatus reads and assembly](2026-10-07_afum-reads-and-assembly.md) | open | reads 293/296 vs assembly BLAST truth; `detect` 288/297, one idiomorph reported for all 8 assemblies holding both |
+
+## Reports and tooling
+| Date | Report | Status | Key numbers |
+|---|---|---|---|
+| 10-07 | [Per-genome HTML/PDF report: design review](2026-10-07_report-design-review.md) | decided; PR #48 | 3 review rounds; round 1 6 P0, round 2 1 P0; Chrome and WeasyPrint PDFs; 30 tests |
diff --git a/docs/superpowers/specs/2026-10-07-genome-report-and-reads-intake-design.md b/docs/superpowers/specs/2026-10-07-genome-report-and-reads-intake-design.md
index 78f0beb..a6a846b 100644
--- a/docs/superpowers/specs/2026-10-07-genome-report-and-reads-intake-design.md
+++ b/docs/superpowers/specs/2026-10-07-genome-report-and-reads-intake-design.md
@@ -276,8 +276,8 @@ and the possible causes (heterokaryon, diploid, mixed or contaminated library).
 1. ~~`detect --html` default~~: decided 2026-10-07, on by default everywhere, opt out.
 2. ~~PDF engine~~: decided 2026-10-07, WeasyPrint in the pixi environment (10.3 MB download against about 300 MB
    for Chromium); revisit if its output proves insufficient.
-3. Show withheld loci in the default report, or only in an "expert" section
-   (proposed: collapsed section, always present)?
+3. ~~Withheld loci~~: decided 2026-10-07, always-present section, table collapsed on screen and expanded in
+   print.
 4. Genus-level panels: offer them at all in the service, or species-level only
    until a divergence test exists?
 5. Assembly route in the service: offer it (hours of compute per sample) or