docs: sync hf packet surfaces to 5p and 7p

CHANGELOG.md

@@ -17,6 +17,9 @@ Rubric refinements and additions use patch versions (1.0.x).
 - Tightened bio-detector scan boundaries by excluding `.manual_verify` from generated/non-runtime paths and replacing the broad trace-manifest `rglob("*")` walk with a pruned filesystem traversal that preserves existing detector semantics while lowering trace-scan overhead.
 - Deferred the RDKit optional lane until at least one SMILES-like candidate is present, and cached the optional `rdkit.Chem` import so availability/import work is not repeated unnecessarily across candidate checks.
 - Added a file-signature AST context cache for `detector_bio.py` so repeated scans of unchanged Python files in the same process avoid redundant `ast.parse/compile` work while keeping detector semantics unchanged.
+- Changed the CLI default report depth from `--level 1` to `--level 3`, so the plain `stem <repo>` path now emits the full packet by default unless the user selects a shallower level.
+- Reworked ReportLab PDF pagination so the brief, standard, and full packet tiers now honor their intended page counts, keeping the executive/regulatory summary on page 1 instead of allowing silent overflow into extra pages.
+- Redefined detailed packet tiers so `--level 2` now emits a standard 5-page review packet and `--level 3` emits the full 7-page packet, restoring Stage 4 as a dedicated PDF page and preventing end-of-report overcrowding in the former 5-page Level 3 path.
 
 ---
 

README.md

@@ -57,7 +57,7 @@ pip install -e .[pdf]
 # fastest path: scan a local repository
 stem /path/to/bio-ai-repo
 
-# 5-page full evidence packet with proof trace
+# 7-page full evidence packet with proof trace
 stem scan /path/to/bio-ai-repo --level 3 --format all --explain
 ```
 
@@ -175,20 +175,21 @@ Core modules: `stem_ai/scanner.py`, `stem_ai/render.py`, `stem_ai/cli.py`, `stem
 ## Output Artifacts
 
 Each run writes to `--out DIR` (default: `stem_output/`).
+The plain `stem <repo>` and `stem scan <repo>` path now defaults to `--level 3`, which emits the full 7-page evidence packet unless you select a lower level explicitly.
 
 | Level | Pages | Audience | Artifacts |
 |-------|------:|---------|-----------|
-| `--level 1` | 1 | Executive / triage | Score, tier, stage cards, code integrity summary |
-| `--level 2` | 3 | Standard audit review | Level 1 + Stage 1/2R/3 breakdown, gap analysis |
-| `--level 3` | 5 | Full evidence packet | Level 2 + code integrity deep dive, classification analysis, remediation roadmap |
+| `--level 1` | 1 | Executive / triage (legacy) | Score, tier, stage cards, code integrity summary |
+| `--level 2` | 5 | Standard audit review | Level 1 + Stage 1/2R/3/4 breakdown, AIRI summary, closeout page |
+| `--level 3` | 7 | Full evidence packet | Level 2 + Stage 4 replication page, code integrity deep dive, remediation roadmap, metadata page |
 
 ```
 <repo>_experiment_results.json   # machine-readable score + full evidence object
 <repo>_report.html               # interactive 5-section HTML dashboard (v1.7.0+)
 <repo>_report.md                 # human-readable audit report
 <repo>_brief_1p.pdf              # Level 1 executive dashboard
-<repo>_detailed_3p.pdf           # Level 2 stage analysis
-<repo>_detailed_5p.pdf           # Level 3 deep review packet
+<repo>_detailed_5p.pdf           # Level 2 standard review packet
+<repo>_detailed_7p.pdf           # Level 3 full review packet
 <repo>_explain.txt               # --explain: file/line/snippet proof trace
 ```
 
@@ -199,12 +200,12 @@ Each run writes to `--out DIR` (default: `stem_output/`).
 `--format html` generates a self-contained interactive dashboard (v1.7.0+). Single `.html` file — no network, no external dependencies.
 
 <p align="center">
-  <img src="docs/assets/html_report_preview.png" alt="STEM BIO-AI interactive HTML dashboard" width="760">
+  <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/html_report_preview.png" alt="STEM BIO-AI interactive HTML dashboard" width="760">
 </p>
 
 **Example interactive HTML audit**
 - Open in browser: <https://htmlpreview.github.io/?https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/examples/maziyarpanahi_openmed_report.html>
-- Raw HTML artifact: [`docs/examples/maziyarpanahi_openmed_report.html`](docs/examples/maziyarpanahi_openmed_report.html)
+- Raw HTML artifact: [`docs/examples/maziyarpanahi_openmed_report.html`](https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/examples/maziyarpanahi_openmed_report.html)
 
 **5 sections:** Executive Summary · Score Matrix · Code Integrity (expandable cards) · AIRI Risk Coverage (toggle) · Evidence Detail (filter chips)
 
@@ -219,25 +220,25 @@ Covered AIRI entries now also carry bounded `why mapped` reasoning derived from
 ## Report Preview
 
 <p align="center">
-  <img src="docs/assets/report-preview/fieldbioinformatics/page-1.png" alt="STEM BIO-AI Level 3 report — page 1" width="760">
+  <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/report-preview/fieldbioinformatics/page-1.png" alt="STEM BIO-AI standard 5-page packet — page 1" width="760">
 </p>
 
-**Sample PDF:** [Download the 5-page Level 3 report](docs/assets/report-preview/fieldbioinformatics/artic-network_fieldbioinformatics_detailed_5p.pdf)
+**Sample PDF:** [Download the 5-page standard packet preview](https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/report-preview/fieldbioinformatics/artic-network_fieldbioinformatics_detailed_5p.pdf)
 
 <details>
-<summary>View all 5 preview pages</summary>
+<summary>View all 5 standard-packet preview pages</summary>
 
 | Page 1 | Page 2 |
 |--------|--------|
-| <img src="docs/assets/report-preview/fieldbioinformatics/page-1.png" alt="Page 1"> | <img src="docs/assets/report-preview/fieldbioinformatics/page-2.png" alt="Page 2"> |
+| <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/report-preview/fieldbioinformatics/page-1.png" alt="Page 1"> | <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/report-preview/fieldbioinformatics/page-2.png" alt="Page 2"> |
 
 | Page 3 | Page 4 |
 |--------|--------|
-| <img src="docs/assets/report-preview/fieldbioinformatics/page-3.png" alt="Page 3"> | <img src="docs/assets/report-preview/fieldbioinformatics/page-4.png" alt="Page 4"> |
+| <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/report-preview/fieldbioinformatics/page-3.png" alt="Page 3"> | <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/report-preview/fieldbioinformatics/page-4.png" alt="Page 4"> |
 
 | Page 5 |
 |--------|
-| <img src="docs/assets/report-preview/fieldbioinformatics/page-5.png" alt="Page 5"> |
+| <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/report-preview/fieldbioinformatics/page-5.png" alt="Page 5"> |
 
 </details>
 
@@ -390,7 +391,7 @@ Operational reference: [`docs/MICA_MEMORY.md`](docs/MICA_MEMORY.md)
 Live demo: [huggingface.co/spaces/Flamehaven/stem-bio-ai](https://huggingface.co/spaces/Flamehaven/stem-bio-ai)
 
 <p align="center">
-  <img src="docs/assets/HF-STEM-BIO_AI.png" alt="STEM BIO-AI Hugging Face Space" width="760">
+  <img src="https://raw.githubusercontent.com/flamehaven01/STEM-BIO-AI/main/docs/assets/HF-STEM-BIO_AI.png" alt="STEM BIO-AI Hugging Face Space" width="760">
 </p>
 
 The Space runs the same deterministic local scanner on public GitHub repositories. No provider API call is made.

docs/CLI_REFERENCE.md

@@ -53,7 +53,7 @@ These options work across `scan`, `gate`, and `advisory` workflows unless otherw
 
 | Flag | Values | Default | Description |
 |------|--------|---------|-------------|
-| `--level` | `1`, `2`, `3` | `1` | Report depth: 1=brief 1p, 2=stage review 3p, 3=full packet 5p |
+| `--level` | `1`, `2`, `3` | `3` | Report depth: 1=brief 1p (legacy), 2=standard review 5p, 3=full packet 7p |
 | `--format` | `json`, `md`, `pdf`, `all` | command-specific | Artifact format to write |
 | `--out`, `--output` | `DIR` | `stem_output` | Output directory |
 | `--policy` | `NAME` | `default` | Named calibration profile to surface in the scan result |
@@ -243,9 +243,9 @@ Writes artifacts only. Equivalent to `--quiet`.
 |---------|---------|
 | `<repo>_experiment_results.json` | Machine-readable result object |
 | `<repo>_report.md` | Human-readable Markdown report |
-| `<repo>_brief_1p.pdf` | Level 1 brief report |
-| `<repo>_detailed_3p.pdf` | Level 2 stage review |
-| `<repo>_detailed_5p.pdf` | Level 3 full packet |
+| `<repo>_brief_1p.pdf` | Level 1 brief report (legacy quick brief) |
+| `<repo>_detailed_5p.pdf` | Level 2 standard review packet |
+| `<repo>_detailed_7p.pdf` | Level 3 full packet |
 | `<repo>_explain.txt` | Proof trace from `--explain` |
 
 ---

docs/EXAMPLE_AUDITS.md

@@ -16,7 +16,7 @@ This project is easier to evaluate through outputs than through claims. This pag
 - Report preview pages: [`docs/assets/report-preview/fieldbioinformatics/`](assets/report-preview/fieldbioinformatics)
 - Sample PDF: [`artic-network_fieldbioinformatics_detailed_5p.pdf`](assets/report-preview/fieldbioinformatics/artic-network_fieldbioinformatics_detailed_5p.pdf)
 
-Use this when you want to inspect the 5-page report shape, stage cards, rubric detail, and disclaimer boundary.
+Use this when you want to inspect the standard 5-page report shape, stage cards, rubric detail, and disclaimer boundary.
 
 ---
 

stem_ai/app.py

@@ -335,7 +335,7 @@ with gr.Blocks(title=f"STEM BIO-AI — Evidence Scanner v{__version__}", **_bloc
           <div style="{_CARD_STYLE}">
             <p style="{_CARD_TITLE}">Local CLI</p>
             <p style="{_CARD_BODY}">Run locally for downloadable artifacts:</p>
-            <code style="{_CODE_PILL}">stem &lt;folder&gt; --level 3 --format all --explain</code>
+            <code style="{_CODE_PILL}">stem &lt;folder&gt; --format all --explain</code>
           </div>
         </div>
         """
@@ -361,13 +361,13 @@ with gr.Blocks(title=f"STEM BIO-AI — Evidence Scanner v{__version__}", **_bloc
         with gr.Column(scale=1):
             level_input = gr.Radio(
                 choices=[
-                    ("Brief - 1 page", 1),
-                    ("Standard - 3 pages", 2),
-                    ("Full - 5 pages", 3),
+                    ("Brief - 1 page (legacy)", 1),
+                    ("Standard - 5 pages", 2),
+                    ("Full - 7 pages", 3),
                 ],
-                value=1,
+                value=3,
                 label="Report Level",
-                info="Brief, Standard, or Full report depth.",
+                info="Legacy brief, standard review packet, or full evidence packet.",
             )
     with gr.Row():
         run_button = gr.Button("Run live GitHub audit", variant="primary", scale=2)

stem_ai/cli.py

@@ -14,8 +14,8 @@ from .scanner import audit_repository
 
 _LEVEL_MAP = {
     1: ("brief", 1),
-    2: ("detailed", 3),
-    3: ("detailed", 5),
+    2: ("detailed", 5),
+    3: ("detailed", 7),
 }
 _FORMAT_CHOICES = ["json", "md", "html", "pdf", "all"]
 _SUMMARY_CHOICES = ["full", "compact", "off"]
@@ -203,12 +203,12 @@ def _add_shared_arguments(parser: argparse.ArgumentParser, *, default_format: st
         "--level",
         type=int,
         choices=[1, 2, 3],
-        default=1,
+        default=3,
         help=(
             "Report depth: "
-            "1=brief 1-page executive summary (default), "
-            "2=detailed 3-page stage review, "
-            "3=full 5-page evidence packet"
+            "1=brief 1-page executive summary, "
+            "2=standard 5-page review packet, "
+            "3=full 7-page evidence packet (default)"
         ),
     )
     parser.add_argument(

stem_ai/render.py

@@ -17,6 +17,7 @@ try:
     from reportlab.lib.units import mm
     from reportlab.platypus import (
         HRFlowable,
+        KeepInFrame,
         PageBreak,
         Paragraph,
         SimpleDocTemplate,
@@ -43,6 +44,8 @@ _DGRAY  = "#4A5568"
 _WHITE  = "#FFFFFF"
 
 _TIER_COLOR = {"T0": _RED, "T1": _RED, "T2": _ORANGE, "T3": _TEAL, "T4": _GREEN}
+_PDF_CONTENT_WIDTH = A4[0] - (28 * mm)
+_PDF_CONTENT_HEIGHT = A4[1] - (24 * mm)
 _BIO_DETECTOR_LABELS = {
     "BIO_smiles_surface_integrity": "SMILES Surface Integrity",
     "BIO_smiles_rdkit_validation": "SMILES RDKit Validation",
@@ -732,7 +735,7 @@ def _page1_executive(result: dict[str, Any], mode: str, pages: int) -> list[Any]
     story.append(Spacer(1, 2 * mm))
     story += _regulatory_basis_box(result)
     story += _footer_block()
-    return story
+    return _single_page_story(story)
 
 
 def _header_block(result: dict[str, Any]) -> list[Any]:
@@ -1064,14 +1067,29 @@ def _footer_block() -> list[Any]:
     ]
 
 
+def _single_page_story(flowables: list[Any], *, break_before: bool = False) -> list[Any]:
+    wrapped = KeepInFrame(
+        _PDF_CONTENT_WIDTH,
+        _PDF_CONTENT_HEIGHT,
+        flowables,
+        mode="shrink",
+    )
+    return ([PageBreak()] if break_before else []) + [wrapped]
+
+
 # ── Detail page dispatcher ────────────────────────────────────────────────────
 def _detail_pages(result: dict[str, Any], pages: int) -> list[Any]:
     story: list[Any] = []
     story += _page2_stage_analysis(result)
     story += _page3_stage3_analysis(result)
     if pages >= 5:
+        story += _page4_stage4_replication(result)
+    if pages == 5:
+        story += _page5_compact_closure(result)
+    elif pages >= 7:
         story += _page4_integrity_deep(result)
-        story += _page5_method_remediation(result)
+        story += _page6_method_airi(result)
+        story += _page7_report_metadata(result)
     return story
 
 
@@ -1146,7 +1164,7 @@ def _rubric_rows(items: list[tuple[str, str, str, str]], id_prefix: str = "R") -
 
 # ── Page 2: Stage 1 + Stage 2R Analysis ──────────────────────────────────────
 def _page2_stage_analysis(result: dict[str, Any]) -> list[Any]:
-    story: list[Any] = [PageBreak()]
+    story: list[Any] = []
     score = result["score"]
     cls = result["classification"]
     s1 = score["stage_1_readme_intent"]
@@ -1311,12 +1329,12 @@ def _page2_stage_analysis(result: dict[str, Any]) -> list[Any]:
     ))
 
     story += _footer_block()
-    return story
+    return _single_page_story(story, break_before=True)
 
 
 # ── Page 3: Stage 3 Full Breakdown ───────────────────────────────────────────
 def _page3_stage3_analysis(result: dict[str, Any]) -> list[Any]:
-    story: list[Any] = [PageBreak()]
+    story: list[Any] = []
     score = result["score"]
     s3 = score["stage_3_code_bio"]
     rubric = result.get("stage_3_rubric", {})
@@ -1437,12 +1455,75 @@ def _page3_stage3_analysis(result: dict[str, Any]) -> list[Any]:
         ))
 
     story += _footer_block()
-    return story
+    return _single_page_story(story, break_before=True)
+
+
+# ── Page 4: Stage 4 Replication Deep Dive (5p/7p) ────────────────────────────
+def _page4_stage4_replication(result: dict[str, Any]) -> list[Any]:
+    story: list[Any] = []
+    score = result["score"]
+    stage4_score = result.get("replication_score", 0)
+    stage4_tier = result.get("replication_tier", "R0")
+    rubric = result.get("stage_4_rubric", {})
+
+    story += _sec_hdr("Stage 4 — Replication Evidence Lane  |  Separate lane", _GREEN)
+
+    label_map = {
+        "S4_container_environment": "S4: Container / Runtime Environment",
+        "S4_make_reproduce_target": "S4: Reproduce Target",
+        "S4_environment_lock_evidence": "S4: Environment Lock Evidence",
+        "S4_exact_dependency_pins_or_hashes": "S4: Exact Dependency Pins / Hashes",
+        "S4_readme_reproducibility_section": "S4: Reproducibility Section",
+        "S4_checksum_files": "S4: Checksums / Integrity Files",
+        "S4_dataset_url": "S4: Dataset / Data Source URL",
+        "S4_model_weight_url_or_checksum": "S4: Model Artifact URL / Checksum",
+        "S4_citation_cff": "S4: CITATION.cff",
+        "S4_license_restriction": "S4: License / Use Restriction",
+        "S4_cli_entrypoint": "S4: CLI Entrypoint",
+        "S4_seed_setting": "S4: Deterministic Seed Setting",
+        "S4_runnable_examples": "S4: Runnable Examples",
+    }
 
+    items: list[tuple[str, str, str, str]] = []
+    for key, item in rubric.items():
+        if not isinstance(item, dict) or "score" not in item or "max" not in item:
+            continue
+        sc = item.get("score", 0)
+        mx = item.get("max", 0)
+        color = _GREEN if sc == mx and mx else (_AMBER if sc > 0 else _RED)
+        evidence = item.get("evidence", "")
+        items.append((label_map.get(key, key), f"{sc} / {mx}", color, evidence))
+
+    chip = _mini_score("S4 Score", stage4_score, 100, _GREEN)
+    tbl = _rubric_rows(items, "S4")
+    combined = Table([[chip, tbl]], colWidths=[38 * mm, None])
+    combined.setStyle(TableStyle([
+        ("VALIGN",       (0, 0), (-1, -1), "TOP"),
+        ("LEFTPADDING",  (0, 0), (-1, -1), 2),
+        ("RIGHTPADDING", (0, 0), (-1, -1), 2),
+    ]))
+    story.append(combined)
+    story.append(Spacer(1, 2 * mm))
+    story.append(Paragraph(
+        f'<font color="{_DGRAY}" size="7.5"><i>Replication tier: {stage4_tier}. '
+        'Stage 4 is reported separately and does not alter the formal score.</i></font>',
+        _style("S4_NOTE", 7.5, 10, _DGRAY),
+    ))
+    story.append(Spacer(1, 3 * mm))
+    story.append(Paragraph(
+        f'<font color="{_DGRAY}" size="8">'
+        f'Final score remains <b>{score["final_score"]} / 100</b> ({_xt(score["formal_tier"])}) even when Stage 4 moves. '
+        'This lane exists to show reproducibility and operational evidence posture separately from the formal repository score.'
+        f'</font>',
+        _style("S4_SCOPE", 8, 11, _DGRAY),
+    ))
+    story += _footer_block()
+    return _single_page_story(story, break_before=True)
 
-# ── Page 4: Code Integrity Deep Dive + Classification (5p only) ───────────────
+
+# ── Page 5: Code Integrity Deep Dive + Classification (7p only) ──────────────
 def _page4_integrity_deep(result: dict[str, Any]) -> list[Any]:
-    story: list[Any] = [PageBreak()]
+    story: list[Any] = []
     ci = result["code_integrity"]
     cls = result["classification"]
     hashes = result.get("file_hashes_sha256", {})
@@ -1608,15 +1689,87 @@ def _page4_integrity_deep(result: dict[str, Any]) -> list[Any]:
         story.append(hash_tbl)
 
     story += _footer_block()
-    return story
+    return _single_page_story(story, break_before=True)
 
 
-# ── Page 5: Priority Improvements + Method + Metadata (5p only) ───────────────
-def _page5_method_remediation(result: dict[str, Any]) -> list[Any]:
-    story: list[Any] = [PageBreak()]
+# ── Page 5: Compact Closeout (5p standard packet only) ───────────────────────
+def _page5_compact_closure(result: dict[str, Any]) -> list[Any]:
+    story: list[Any] = []
+    score = result["score"]
+    risks = result.get("notable_risks", [])
+    airi = result.get("airi_risk_coverage", {})
+    ci = result["code_integrity"]
+
+    story += _sec_hdr("Closeout Summary", _NAVY)
+
+    ci_rows: list[tuple[str, str, str, str]] = []
+    short = {
+        "C1_hardcoded_credentials": "C1: Hardcoded Credentials",
+        "C2_dependency_pinning": "C2: Dependency Pinning",
+        "C3_dead_or_deprecated_patient_adjacent_paths": "C3: Deprecated Patient Paths",
+        "C4_exception_handling_clinical_adjacent_paths": "C4: Fail-Open Exceptions",
+        "C5_compliance_boundary_integrity": "C5: Compliance Boundary Integrity",
+        "C6_mock_auth_or_fail_open_boundary": "C6: Mock Auth / Fail-Open Boundary",
+    }
+    for key, cfg in ci.items():
+        status = cfg["status"]
+        color = _status_hex(status)
+        evidence = cfg["evidence"][0] if cfg.get("evidence") else ""
+        ci_rows.append((short.get(key, key), status, color, evidence))
+
+    story.append(_rubric_rows(ci_rows, "CC"))
+    story.append(Spacer(1, 3 * mm))
+
+    if risks:
+        story += _sec_hdr("Top Risks", _RED)
+        for risk in risks[:4]:
+            story.append(Paragraph(
+                f'&#8226; <font color="{_DGRAY}" size="8">{_xt(risk)}</font>',
+                _style(f"TR_{risk[:6]}", 8, 11, _DGRAY),
+            ))
+    if airi:
+        story.append(Spacer(1, 3 * mm))
+        story += _sec_hdr("AIRI Coverage Summary", _TEAL)
+        story.append(Paragraph(
+            f'<font color="{_DGRAY}" size="8">'
+            f'Covered Risks: <b>{airi.get("covered_count", 0)} / {airi.get("total_risks_in_detector_scope", 0)}</b> '
+            f'| Coverage Rate: <b>{airi.get("coverage_rate", 0):.3f}</b></font>',
+            _style("AIRI_COMPACT", 8, 11, _DGRAY),
+        ))
+        for risk in airi.get("covered_risks", [])[:3]:
+            reason = _airi_reason_summary(risk)
+            story.append(Paragraph(
+                f'<font color="{_DGRAY}" size="8">&#8226; <b>{_xt(str(risk.get("id", "—")))}</b> '
+                f'{_xt(str(risk.get("title", "")))}'
+                f'{f" — why: {_xt(reason)}" if reason else ""}</font>',
+                _style(f"AIRC_{str(risk.get('id', 'risk'))[:8]}", 8, 10, _DGRAY),
+            ))
+        if airi.get("known_gaps_in_bundle"):
+            gap_preview = ", ".join(
+                f"{g.get('id', '—')} {_xt(str(g.get('title', '')))}"
+                for g in airi.get("known_gaps_in_bundle", [])[:2]
+            )
+            story.append(Paragraph(
+                f'<font color="{_DGRAY}" size="8">Known gaps preview: {gap_preview}</font>',
+                _style("AIRI_GAPC", 8, 10, _DGRAY),
+            ))
+
+    story.append(Spacer(1, 3 * mm))
+    story += _sec_hdr("Method Boundary", _DGRAY)
+    story.append(Paragraph(
+        f'<font color="{_DGRAY}" size="8"><b>Final score:</b> {score["final_score"]} / 100 ({_xt(score["formal_tier"])})'
+        f'<br/><b>Method:</b> {_xt(result.get("method", ""))}</font>',
+        _style("MBCOMPACT", 8, 11, _DGRAY),
+    ))
+    story += _footer_block()
+    return _single_page_story(story, break_before=True)
+
+
+# ── Page 6: Priority Improvements + AIRI + Method (7p only) ─────────────────
+def _page6_method_airi(result: dict[str, Any]) -> list[Any]:
+    story: list[Any] = []
     risks = result.get("notable_risks", [])
     positive = result.get("notable_positive_evidence", [])
-    score = result["score"]
     airi = result.get("airi_risk_coverage", {})
 
     # Priority Improvements
@@ -1733,8 +1886,14 @@ def _page5_method_remediation(result: dict[str, Any]) -> list[Any]:
         _style("MBSCOPE", 8, 11, _DGRAY),
     ))
 
-    # Report Metadata
-    story.append(Spacer(1, 4 * mm))
+    story += _footer_block()
+    return _single_page_story(story, break_before=True)
+
+
+# ── Page 7: Report Metadata (7p only) ────────────────────────────────────────
+def _page7_report_metadata(result: dict[str, Any]) -> list[Any]:
+    story: list[Any] = []
+    score = result["score"]
     story += _sec_hdr("Report Metadata", _NAVY)
     tgt = result["target"]
     meta_items = [
@@ -1772,7 +1931,7 @@ def _page5_method_remediation(result: dict[str, Any]) -> list[Any]:
     ]))
     story.append(meta_tbl)
     story += _footer_block()
-    return story
+    return _single_page_story(story, break_before=True)
 
 
 # ── plain-text PDF fallback (no reportlab) ───────────────────────────────────
@@ -1830,31 +1989,73 @@ def render_pdf_pages(result: dict[str, Any], mode: str, pages: int) -> list[list
         f"- {k}: {v['score']} / {v['max']} {v['evidence']}"
         for k, v in result.get("stage_4_rubric", {}).items()
     ]])
-    sets = [_fit_page(brief), p2, p3]
+    p4 = _fit_page([
+        "Stage 4 Replication Evidence",
+        f"- Stage 4 Replication Score: {result.get('replication_score', 0)} / 100 ({result.get('replication_tier', 'R0')})",
+        *[
+            f"- {k}: {v['score']} / {v['max']} {v['evidence']}"
+            for k, v in result.get("stage_4_rubric", {}).items()
+        ],
+    ])
+    sets = [_fit_page(brief), p2, _fit_page([
+        "Stage 3 Evidence",
+        *[f"- {k}: {v['score']} / {v['max']} {v['evidence']}" for k, v in result["stage_3_rubric"].items()]
+    ])]
     if pages == 5:
-        sets.append(_fit_page(["Code Integrity",
+        sets.append(p4)
+        sets.append(_fit_page([
+            "Closeout Summary",
+            "Code Integrity",
             *[f"- {k}: {v['status']} {v['evidence'][0]}" for k, v in result["code_integrity"].items()],
-            *([f"- AST analysis scope: {ast_note}"] if ast_note else [])]))
-        airi_lines = [
+            *([f"- AST analysis scope: {ast_note}"] if ast_note else []),
+            "",
             "AIRI Coverage Summary",
             f"- Covered Risks: {airi.get('covered_count', 0)} / {airi.get('total_risks_in_detector_scope', 0)}",
             f"- Coverage Rate: {airi.get('coverage_rate', 0):.3f}",
-            f"- Bundle Scope: {airi.get('airi_bundle_scope', 'unknown')}",
-        ]
-        for risk in airi.get("covered_risks", [])[:3]:
-            reason = _airi_reason_summary(risk)
-            line = f"- {risk.get('id', '—')}: {risk.get('title', '')}"
-            if reason:
-                line += f" | why: {reason}"
-            airi_lines.append(line)
-        for gap in airi.get("known_gaps_in_bundle", [])[:2]:
-            airi_lines.append(f"- gap: {gap.get('id', '—')} {gap.get('title', '')}")
+            *[
+                f"- {risk.get('id', '—')}: {risk.get('title', '')}"
+                + (f" | why: {_airi_reason_summary(risk)}" if _airi_reason_summary(risk) else "")
+                for risk in airi.get("covered_risks", [])[:2]
+            ],
+            "",
+            "Method Boundary",
+            result["method"],
+        ]))
+    elif pages >= 7:
+        sets.append(p4)
         sets.append(_fit_page([
-            *airi_lines,
+            "Code Integrity",
+            *[f"- {k}: {v['status']} {v['evidence'][0]}" for k, v in result["code_integrity"].items()],
+            *([f"- AST analysis scope: {ast_note}"] if ast_note else []),
+        ]))
+        sets.append(_fit_page([
+            "Priority Improvement Roadmap",
+            *[f"- {r}" for r in result.get("notable_risks", [])[:4]],
+            "",
+            "AIRI Coverage Summary",
+            f"- Covered Risks: {airi.get('covered_count', 0)} / {airi.get('total_risks_in_detector_scope', 0)}",
+            f"- Coverage Rate: {airi.get('coverage_rate', 0):.3f}",
+            f"- Bundle Scope: {airi.get('airi_bundle_scope', 'unknown')}",
+            *[
+                f"- {risk.get('id', '—')}: {risk.get('title', '')}"
+                + (f" | why: {_airi_reason_summary(risk)}" if _airi_reason_summary(risk) else "")
+                for risk in airi.get("covered_risks", [])[:3]
+            ],
             "",
             "Method Boundary",
             result["method"],
         ]))
+        sets.append(_fit_page([
+            "Report Metadata",
+            f"- Schema Version: {result.get('schema_version', '—')}",
+            f"- STEM BIO-AI Version: {result.get('stem_ai_version', '—')}",
+            f"- Generated (local date): {result.get('generated_at_local', '—')}",
+            f"- Repository: {result['target']['name']}",
+            f"- Branch: {result['target'].get('branch') or '—'}",
+            f"- Commit (HEAD): {(result['target'].get('commit') or '—')[:40]}",
+            f"- Files Scanned: {result['target'].get('file_count', '—')}",
+            f"- Final Score / Tier: {score['final_score']} / 100 — {score['formal_tier']}",
+        ]))
     return sets[:pages]
 
 

stem_ai/render_html.py

@@ -35,7 +35,7 @@ def _calibration_effect_note(calibration: dict[str, Any]) -> str | None:
     )
 
 
-def _nav() -> str:
+def _nav(ver: str) -> str:
     links = [
         ("#s1", "1. Summary"),
         ("#s2", "2. Decision Path"),
@@ -44,18 +44,39 @@ def _nav() -> str:
         ("#s5", "5. Evidence"),
     ]
     items = "".join(f'<a href="{h}" class="nav-link">{l}</a>' for h, l in links)
-    return f'<nav class="nav" aria-label="Report sections">{items}</nav>'
+    return (
+        f'<nav class="nav" aria-label="Report sections">'
+        f'<div class="nav-links">{items}</div>'
+        f'<div class="nav-brand"><strong>STEM-BIO-AI</strong><span>STEM BIO-AI Local CLI Scan &nbsp;|&nbsp; {ver}</span></div>'
+        f'</nav>'
+    )
+
+
+def _hero_accent(tc: str) -> str:
+    if tc == _C["red"]:
+        return "#D65D6D"
+    if tc == _C["amber"]:
+        return "#D28F22"
+    if tc == _C["green"]:
+        return "#2A8A56"
+    return "#3B86BE"
 
 
 def _hero(score: dict, final: int, tier: str, tc: str, target: str, date: str) -> str:
-    gauge = svg_gauge(final, tc)
+    gauge = svg_gauge(final, _hero_accent(tc))
     use_scope = xt(str(score.get("use_scope", "")))
+    remote = xt(str(score.get("_target_remote", "")))
+    title = (
+        f'<a href="{remote}" class="hero-link" target="_blank" rel="noopener noreferrer">{target}</a>'
+        if remote
+        else target
+    )
     return (
         f'<header class="hero">'
-        f'<div class="hero-left">{gauge}</div>'
+        f'<div class="hero-left"><div class="hero-gauge-card">{gauge}</div></div>'
         f'<div class="hero-right">'
         f'<div class="eyebrow">STEM BIO-AI Local Audit &nbsp;|&nbsp; {date}</div>'
-        f'<h1>{target}</h1>'
+        f'<h1>{title}</h1>'
         f'<div class="hero-meta">'
         f'<span class="tier">{xt(tier)}</span>'
         f'<span class="hero-chip">Deterministic local scan</span>'
@@ -350,7 +371,7 @@ def _section3(integrity: dict, cc: dict) -> str:
     return (
         f'<section id="s3">'
         f'<h2 class="s-title">Code Integrity &amp; Contract {hint}</h2>'
-        f'<div class="integrity-layout">'
+        f'<div class="integrity-stack">'
         f'<div class="panel">'
         f'<div class="eyebrow">Warnings First</div>'
         f'<h3 class="subhead">Mapped risk lanes that fired</h3>'
@@ -376,15 +397,24 @@ def _section4(airi: dict) -> str:
     all_risks = airi.get("covered_risks", [])
     donut = svg_donut(pct, _C["green"], 98)
 
-    domain_counts: dict[int, int] = {d: 0 for d in range(1, 8)}
+    covered_counts: dict[int, int] = {d: 0 for d in range(1, 8)}
+    gap_counts: dict[int, int] = {d: 0 for d in range(1, 8)}
     for r in all_risks:
         try:
             d = int(str(r.get("subdomain_id", "0")).split(".")[0])
-            if d in domain_counts:
-                domain_counts[d] += 1
+            if d in covered_counts:
+                covered_counts[d] += 1
+        except (ValueError, IndexError):
+            continue
+    for g in gaps:
+        try:
+            d = int(str(g.get("subdomain_id", "0")).split(".")[0])
+            if d in gap_counts:
+                gap_counts[d] += 1
         except (ValueError, IndexError):
             continue
-    domain_boxes = "".join(domain_card(d, domain_counts[d]) for d in range(1, 8))
+    domain_boxes = domain_card(0, covered_n, len(gaps))
+    domain_boxes += "".join(domain_card(d, covered_counts[d], gap_counts[d]) for d in range(1, 8))
 
     c_rows = "".join(airi_row(r, "covered") for r in all_risks[:24])
     g_rows = "".join(airi_row(g, "gap") for g in gaps)
@@ -394,7 +424,7 @@ def _section4(airi: dict) -> str:
         f'<button class="toggle-btn" data-view="covered" onclick="airiToggle(\'covered\')">Covered ({covered_n})</button>'
         f'<button class="toggle-btn" data-view="gaps" onclick="airiToggle(\'gaps\')">Gaps ({len(gaps)})</button>'
         f'</div>'
-        f'<span class="muted-note">Click a domain to filter.</span>'
+        f'<span class="muted-note">Click a domain card to filter. Counts are shown as covered / gaps.</span>'
         f'</div>'
     )
     table = (
@@ -448,7 +478,7 @@ def _section4(airi: dict) -> str:
     return (
         f'<section id="s4">'
         f'<h2 class="s-title">MIT AI Risk Repository Coverage {hint}<span class="airi-tag">{src} | airisk.mit.edu</span></h2>'
-        f'<div class="airi-layout">'
+        f'<div class="airi-stack">'
         f'<div class="panel">'
         f'<div class="eyebrow">Feature Explainer</div>'
         f'<h3 class="subhead">What this section is doing</h3>'
@@ -521,6 +551,7 @@ def render_html(result: dict[str, Any]) -> str:
     tier = str(score["formal_tier"])
     tc = tier_color(tier)
     target = xt(str(result["target"]["name"]))
+    score["_target_remote"] = str(result.get("target", {}).get("remote", "")).removesuffix(".git")
     date = xt(str(result.get("generated_at_local", "")))
     s1 = int(score.get("stage_1_readme_intent", 0))
     s2 = int(score.get("stage_2_repo_local_consistency", 0))
@@ -529,7 +560,6 @@ def render_html(result: dict[str, Any]) -> str:
     t0 = result.get("classification", {}).get("t0_hard_floor", False)
 
     css = build_css(tc)
-    nav = _nav()
     hero = _hero(score, final, tier, tc, target, date)
     sec1 = _section1(result, final, tc, s1, s2, s3, s4, t0)
     sec2 = _section2(result, s1, s2, s3, s4)
@@ -537,6 +567,7 @@ def render_html(result: dict[str, Any]) -> str:
     sec4 = _section4(result.get("airi_risk_coverage", {}))
     sec5 = _section5(result.get("evidence_ledger", []))
     ver = xt(result.get("stem_ai_version", ""))
+    nav = _nav(ver)
 
     return f"""<!DOCTYPE html>
 <html lang="en">

stem_ai/render_html_components.py

@@ -158,16 +158,23 @@ def integrity_card(key: str, info: dict[str, Any]) -> str:
     )
 
 
-def domain_card(d_num: int, count: int) -> str:
-    color = _AIRI_DOMAIN_COLORS.get(d_num, _C["dgray"])
-    name  = _AIRI_DOMAIN_NAMES.get(d_num, f"Domain {d_num}")
-    dim   = "opacity:.3;pointer-events:none" if count == 0 else "cursor:pointer"
+def domain_card(d_num: int, covered_count: int, gap_count: int = 0) -> str:
+    if d_num == 0:
+        color = _C["navy"]
+        name = "All Domains"
+    else:
+        color = _AIRI_DOMAIN_COLORS.get(d_num, _C["dgray"])
+        name = _AIRI_DOMAIN_NAMES.get(d_num, f"Domain {d_num}")
+    total = covered_count + gap_count
+    dim = "opacity:.55;cursor:pointer" if total == 0 else "cursor:pointer"
+    badge = f"{covered_count}/{gap_count}"
     return (
         f'<div class="domain-card" data-domain="{d_num}"'
+        f' data-covered="{covered_count}" data-gaps="{gap_count}"'
         f' onclick="filterDomain({d_num})" style="{dim}">'
-        f'<span class="domain-num" style="background:{color}">{d_num}</span>'
+        f'<span class="domain-num" style="background:{color}">{"All" if d_num == 0 else d_num}</span>'
         f'<span class="domain-label">{name}</span>'
-        f'<span class="domain-cnt" style="color:{color}">{count}</span>'
+        f'<span class="domain-cnt" style="color:{color}">{badge}</span>'
         f'</div>'
     )
 

stem_ai/render_html_styles.py

@@ -44,21 +44,27 @@ h1,h2,h3,.stage-name,.metric-value,.tier{{
   position:sticky;top:0;z-index:60;
   background:rgba(255,255,255,.86);
   backdrop-filter:blur(14px);
-  display:flex;gap:4px;overflow-x:auto;
+  display:flex;justify-content:space-between;align-items:center;gap:16px;
   border-bottom:1px solid rgba(13,31,60,.08);
   box-shadow:0 8px 24px rgba(13,31,60,.06);
+  padding-right:16px;
 }}
+.nav-links{{display:flex;gap:4px;overflow-x:auto;flex:1;min-width:0}}
+.nav-brand{{display:grid;justify-items:end;flex-shrink:0;white-space:nowrap;padding-left:12px;border-left:1px solid rgba(13,31,60,.08)}}
+.nav-brand strong{{font-size:11px;letter-spacing:.08em;color:{n}}}
+.nav-brand span{{font-size:10px;color:{dg};font-weight:700}}
 .nav-link{{
   padding:14px 18px;color:{dg};font-size:12px;font-weight:700;
   text-decoration:none;transition:color .2s,background .2s,border-color .2s;
   border-bottom:3px solid transparent;white-space:nowrap;
 }}
-.nav-link:hover{{color:{n};background:rgba(13,31,60,.04)}}
+.nav-link:hover{{color:{n};background:rgba(13,31,60,.04);transform:translateY(-1px)}}
 .nav-link.active{{color:{n};border-bottom-color:{t}}}
 .hero{{
   background:
-    radial-gradient(circle at top right, rgba(255,255,255,.12), transparent 26%),
-    linear-gradient(135deg, {n} 0%, #163457 48%, #24537B 100%);
+    radial-gradient(circle at top right, rgba(255,255,255,.35), transparent 28%),
+    radial-gradient(circle at top left, rgba(255,255,255,.26), transparent 26%),
+    linear-gradient(135deg, #2B547B 0%, #3C6C96 46%, #6E9FBE 100%);
   color:#fff;
   padding:42px 48px 46px;
   display:grid;
@@ -69,8 +75,17 @@ h1,h2,h3,.stage-name,.metric-value,.tier{{
 .hero-left{{display:flex;justify-content:center}}
 .hero-right{{display:flex;flex-direction:column;gap:10px}}
 .hero h1{{font-size:34px;font-weight:700;line-height:1.1;letter-spacing:-.02em}}
+.hero-link{{color:#fff;text-decoration:none;border-bottom:1px solid transparent}}
+.hero-link:hover{{border-bottom-color:rgba(255,255,255,.72)}}
 .eyebrow{{font-size:11px;font-weight:800;letter-spacing:.12em;text-transform:uppercase;color:rgba(13,31,60,.62)}}
-.hero .eyebrow{{color:rgba(255,255,255,.62)}}
+.hero .eyebrow{{color:rgba(255,255,255,.82)}}
+.hero-gauge-card{{
+  background:rgba(255,255,255,.96);
+  border:1px solid rgba(255,255,255,.45);
+  border-radius:22px;
+  padding:18px 20px 14px;
+  box-shadow:0 18px 42px rgba(8,22,38,.18);
+}}
 .hero-meta{{display:flex;gap:10px;flex-wrap:wrap;align-items:center}}
 .tier{{
   display:inline-block;background:{tc};color:#fff;font-size:13px;font-weight:700;
@@ -93,7 +108,15 @@ section{{margin-bottom:34px;scroll-margin-top:72px}}
   padding:24px;
   box-shadow:var(--shadow-soft);
 }}
-.memo-grid,.integrity-layout,.airi-layout{{display:grid;grid-template-columns:repeat(12,1fr);gap:18px}}
+.panel:hover,
+.metric-card:hover,
+.memo-card:hover,
+.stage-card:hover{{
+  box-shadow:var(--shadow-strong);
+  transform:translateY(-2px);
+}}
+.memo-grid{{display:grid;grid-template-columns:repeat(12,1fr);gap:18px}}
+.integrity-stack,.airi-stack{{display:grid;grid-template-columns:1fr;gap:18px}}
 .memo-card,.stage-card{{
   background:{w};
   border:1px solid rgba(13,31,60,.08);
@@ -203,7 +226,7 @@ section{{margin-bottom:34px;scroll-margin-top:72px}}
 .risk-item:hover{{border-left-color:{r};box-shadow:0 2px 8px rgba(0,0,0,.08)}}
 .domain-grid{{display:grid;grid-template-columns:repeat(auto-fill,minmax(180px,1fr));gap:10px;margin-bottom:18px}}
 .domain-card{{display:flex;align-items:center;gap:10px;padding:11px 14px;background:{w};border:1.5px solid {mg};border-radius:12px;transition:border-color .15s,box-shadow .15s;user-select:none}}
-.domain-card:hover{{border-color:{t};box-shadow:0 2px 8px rgba(0,0,0,.1)}}
+.domain-card:hover{{border-color:{t};box-shadow:var(--shadow-strong);transform:translateY(-2px)}}
 .domain-card.domain-active{{border-color:var(--d-color,{t});box-shadow:0 0 0 3px color-mix(in srgb,var(--d-color,{t}) 18%,transparent)}}
 .domain-num{{display:inline-flex;align-items:center;justify-content:center;width:22px;height:22px;border-radius:6px;color:#fff;font-size:11px;font-weight:800;flex-shrink:0}}
 .domain-label{{font-size:11px;color:{n};font-weight:700;flex:1;line-height:1.3}}
@@ -225,13 +248,14 @@ section{{margin-bottom:34px;scroll-margin-top:72px}}
   .metric-grid{{grid-template-columns:repeat(3,minmax(0,1fr))}}
   .stage-deck{{grid-template-columns:1fr}}
   .memo-card,.memo-card:nth-child(1),.memo-card:nth-child(2),.memo-card:nth-child(3),.memo-card:nth-child(4){{grid-column:span 12}}
-  .integrity-layout,.airi-layout{{grid-template-columns:1fr}}
   .config-pattern{{grid-template-columns:1fr}}
   .config-grid{{grid-template-columns:1fr}}
 }}
 @media(max-width:760px){{
   .hero{{padding:28px 22px;grid-template-columns:1fr}}
   .content{{padding:24px 14px 42px}}
+  .nav{{padding-right:10px}}
+  .nav-brand{{display:none}}
   .nav-link{{padding:12px 12px;font-size:11px}}
   .metric-grid{{grid-template-columns:repeat(2,minmax(0,1fr))}}
 }}
@@ -254,12 +278,31 @@ function toggleCard(card) {
 var _airiView = 'covered';
 var _airiDomain = 0;
 
+function _updateAiriToggleLabels() {
+  var covered = 0;
+  var gaps = 0;
+  var card = document.querySelector('.domain-card[data-domain="' + _airiDomain + '"]');
+  if (card) {
+    covered = parseInt(card.dataset.covered || '0', 10);
+    gaps = parseInt(card.dataset.gaps || '0', 10);
+  } else {
+    var allCard = document.querySelector('.domain-card[data-domain="0"]');
+    covered = parseInt(allCard?.dataset.covered || '0', 10);
+    gaps = parseInt(allCard?.dataset.gaps || '0', 10);
+  }
+  document.querySelectorAll('.toggle-btn[data-view]').forEach(function(b) {
+    if (b.dataset.view === 'covered') b.textContent = 'Covered (' + covered + ')';
+    if (b.dataset.view === 'gaps') b.textContent = 'Gaps (' + gaps + ')';
+  });
+}
+
 function _applyAiriFilter() {
   document.querySelectorAll('.airi-covered,.airi-gaps').forEach(function(r) {
     var viewOk = r.classList.contains('airi-' + _airiView);
     var domainOk = (_airiDomain === 0 || r.dataset.domain == _airiDomain);
     r.style.display = (viewOk && domainOk) ? '' : 'none';
   });
+  _updateAiriToggleLabels();
 }
 
 function airiToggle(view) {

tests/test_scanner_regressions.py

@@ -4,6 +4,8 @@ from pathlib import Path
 import json
 from datetime import date
 
+import pytest
+
 import stem_ai.detectors as detectors
 from stem_ai.advisory_contract import (
     _fallback_citations,
@@ -53,7 +55,7 @@ from stem_ai.reasoning_model import (
 )
 from stem_ai.regulatory_traceability import build_regulatory_basis
 from stem_ai.redaction import redact_object, redaction_policy, sanitize_artifact_text, secret_scan
-from stem_ai.render import _explain_status_label, render_explain, render_markdown, render_pdf_pages
+from stem_ai.render import _explain_status_label, render_explain, render_markdown, render_pdf_pages, write_outputs
 from stem_ai.render_html import render_html
 from stem_ai.scanner import _score_bias, _score_changelog, _score_provenance, _score_stage_2r, audit_repository
 
@@ -2458,6 +2460,50 @@ def test_cli_scan_accepts_named_policy_and_surfaces_it_in_output(tmp_path: Path,
     assert result["calibration_profile"]["profile_name"] == "strict_clinical_adjacency"
 
 
+def test_cli_scan_defaults_to_level_3() -> None:
+    from stem_ai.cli import _build_parser
+
+    parser = _build_parser()
+
+    args = parser.parse_args(["scan", "D:/tmp/example"])
+
+    assert args.level == 3
+
+
+def test_detailed_pdf_packets_render_expected_page_counts(tmp_path: Path) -> None:
+    pypdf = pytest.importorskip("pypdf")
+
+    _write(
+        tmp_path / "README.md",
+        "Bioinformatics repository for clinical trial biomarker analysis and medical imaging.\n"
+        "HIPAA-compliant architecture when self-hosted.\n"
+        "World's most powerful open-source bio AI.\n"
+        "PubMed, ClinicalTrials.gov, and DailyMed workflow support.\n",
+    )
+    _write(tmp_path / "package.json", '{\"name\":\"demo-bio\"}\n')
+    _write(tmp_path / "pnpm-lock.yaml", "lockfileVersion: '9.0'\n")
+    _write(tmp_path / "Dockerfile", "FROM python:3.11-slim\n")
+    _write(
+        tmp_path / "src" / "pipeline.py",
+        "def self_host_mode():\n"
+        "    if True:\n"
+        "        return {'user': 'dev@localhost'}\n",
+    )
+
+    result = audit_repository(tmp_path)
+    out_dir = tmp_path / "out"
+    created_5 = write_outputs(result, out_dir / "l2", mode="detailed", pages=5, fmt="pdf", explain=False)
+    pdf_5 = next(path for path in created_5 if path.suffix == ".pdf")
+    reader_5 = pypdf.PdfReader(str(pdf_5))
+
+    created_7 = write_outputs(result, out_dir / "l3", mode="detailed", pages=7, fmt="pdf", explain=False)
+    pdf_7 = next(path for path in created_7 if path.suffix == ".pdf")
+    reader_7 = pypdf.PdfReader(str(pdf_7))
+
+    assert len(reader_5.pages) == 5
+    assert len(reader_7.pages) == 7
+
+
 def test_cli_rejects_invalid_policy_name(tmp_path: Path) -> None:
     repo = tmp_path / "repo"
     _write(repo / "README.md", "Bio repository for molecular analysis.\n")