feat(lenses): Leviticus 11-20 lens content, batch 2 of Pentateuch-rest pilot (#820, #1782)

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s

[CraigBuckmaster]

Summary

Batch 2 of the Pentateuch-rest pilot tracked under #1782. Adds 33 hermeneutic-lens entries across Leviticus 11–20, picking up after #1801 (lev1–10).

After this PR, the Lev pilot has 7 chapters left (lev21–lev27) — that's likely batch 3.

Per-chapter distribution (33 total)

Chapter	Lenses	Why
lev11 (clean/unclean)	3	canonical, redemptive, typological
lev12 (childbirth purity)	2	sparse — christocentric, canonical
lev13 (skin disease pt 1)	2	procedural — grammatical, literary
lev14 (cleansing + house mildew)	3	typological, christocentric, canonical
lev15 (discharges)	2	sparse — christocentric, redemptive
lev16 (Day of Atonement)	6	theology peak — full lens spread
lev17 (life-blood + central altar)	4	christocentric, canonical, literary, redemptive
lev18 (sexual ethics)	3	grammatical, canonical, redemptive
lev19 (Holiness Code opener)	5	theology peak — literary inclusio, mission, devotional, canonical, redemptive
lev20 (capital cases)	3	literary, canonical, redemptive

Pattern: average 3.3 lenses/chapter across the batch, with the two theological peaks (lev16, lev19) maxed and the procedural chapters (lev12, lev13, lev15) pared back. Mirrors PR #1801's curated-sparse strategy.

Theological highlights

• lev16 (Day of Atonement, 6 lenses): Hebrew kaphar / kapporet wordplay grammatically; the two-goat paired type prefiguring Christ's single act (Heb 9:6–14); once-a-year veil entry (v 2) framing Heb 9:11–12; the canonical lexicon Hebrews adopts wholesale; v 29's annual self-denial as devotional rhythm; vv 29–34's perpetual statute as the redemptive arc compressed.
• lev17: v 11's "the life of a creature is in the blood" anchors christocentric (Heb 9:22 cites it directly), canonical (thread to Mt 26:28), literary (vv 1–9 / 10–16 doublet), and redemptive (one altar, one sacrifice).
• lev19 (Holiness Code, 5 lenses): vv 2/37 inclusio for literary; vv 33–34's love-of-foreigner as outward mission posture; vv 9–10 gleaning as devotional release; v 18's love-of-neighbour echoing across six NT citations (Mt/Mk/Lk/Rom/Gal/Jas) for canonical; vv 35–36 honest weights for redemptive marketplace ethics.
• lev14: vv 4–7 two-bird rite as resurrection-pattern type; the Mk 1:40–45 leper Christ sent to fulfil this very rite.
• lev18 + lev20 land-vomit motif: twinned redemptive lenses tying penalty to covenant promise.

Pipeline gate results

schema_validator.py 145332 passed, 0 failed, 19 warnings (pre-existing ESV downloadable warns)
lens_quality_scorer.py 33/33 entries at 100/100 (per-chapter run, floor=100)
build_sqlite.py scripture.db: 102MB, hermeneutic_lenses chapter rows = 442
(= 409 baseline from #1801 + 33 batch 2; matches expected)
validate_sqlite.py 101 passed, 0 failed, 2 warnings (pre-existing embeddings/prompts)

Per-chapter SQLite distribution post-build (chapter_lens_content):

lev11: 3 lev12: 2 lev13: 2 lev14: 3 lev15: 2
lev16: 6 lev17: 4 lev18: 3 lev19: 5 lev20: 3

Plagiarism guards applied

The brief flagged three chapters as iconic-saying-heavy. Anchors deliberately steer off the famous verses:

• lev16 (DoA): entries anchor on kaphar/kapporet Hebrew (grammatical), the paired-goat type (typological), v 2 once-a-year entry (christocentric), the Hebrews vocabulary (canonical), v 29 self-denial (devotional), vv 29–34 perpetual statute (redemptive). The "scapegoat carries sins away" phrasing is avoided throughout.
• lev19 (Holiness Code): entries anchor on the structural inclusio (vv 2/37 for the literary lens), the foreigner law (vv 33–34, mission), gleaning laws (vv 9–10, devotional), honest weights (vv 35–36, redemptive). v 18's "love your neighbour" appears once, in the canonical lens, used as a cross-reference launcher into six NT citations rather than as a stand-alone restatement.
• lev20 (capital cases): entries anchor on the structural penalty-list-then-exhortation shape (literary), v 7's consecration call (canonical), and vv 22–23 land-vomit (redemptive). Molech (vv 2–5) and the separation formula (vv 24–26) are deliberately not the anchor verses.

Themes-panel plagiarism: the themes panel for these chapters is structurally thin (label/score array + a one-line title), so the 60% word-overlap check in the scorer poses minimal risk in practice. Verified by 100/100 per-chapter scores including the relevance dimension.

Filler-pattern guards

All six banned phrases (the lens reveals, this lens shows, through this lens, this passage shows that, this text teaches that, in this chapter we see/learn) verified absent across the 33 entries by an authoring-time regex sweep.

Rubric-token trap watch (carried forward from batch 1)

The lens-name-substring trap was the chief failure mode in #1801. Every entry was authored against the rubric's actual token list (lens_rubrics.json) rather than relying on the lens name. Specifically verified:

• typological entries use type / pattern / prefigures / foreshadow — never just "typological".
• canonical entries use canon / canonical / echoes / thread / across Scripture — never just "canonical".
• christocentric entries name Christ / Jesus / Messiah / gospel explicitly.
• mission entries use mission / nations / outward (lev19 only).

Length discipline

All entries authored to a 250-char target with 30-char headroom against the 280-char schema ceiling. Actual range: 200–228 chars. Median ~215.

Watch list for tier-2 audit

These are the entries most worth a human spot-check during accuracy auditing:

1. lev14 typological — the resurrection-pattern reading of the two-bird rite is a defensible (Calvin and Bonar both develop it) but not universal interpretation. Anchor verse vv 4–7 is solid; the "pattern prefigures resurrection" gloss is the interpretive move.
2. lev16 typological — "Heb 9:6–14 reads the pair as a single anticipation" is a strong claim. Verify the citation supports a single-act reading rather than a sequential one.
3. lev17 christocentric — "Heb 9:22 cites this verse directly" — Heb 9:22 echoes the principle rather than quoting Lev 17:11 verbatim. The phrasing "cites this verse" is intentional shorthand for quotation-of-principle but could be tightened to "echoes this verse" if a tier-2 reviewer prefers stricter language.
4. lev19 canonical — the six-citation roster (Mt 22:39, Mk 12:31, Lk 10:27, Rom 13:9, Gal 5:14, Jas 2:8) is a known set; double-check none has been miscredited.

Out of scope

• No new content generation in content/leviticus/{N}.json itself (only the lens chapter files).
• No app code changes.
• No CI workflow changes.
• scripture.db rebuild is local-only — production R2 upload happens via _tools/upload_to_r2.py post-merge.
• app/assets/db-manifest.json and app/assets/explore-images.json drift was checked out before staging, per the brief's git checkout -- step.

Rollback

git revert <merge-commit> is sufficient. No schema changes, no migration, no R2 mutation in this PR.

Refs

• Epic: Epic: Hermeneutic Lenses Content #820 (Hermeneutic Lenses)
• Tracker: [#820 / 3] Pentateuch rest — Exodus, Leviticus, Numbers, Deuteronomy (137 ch) #1782 (Pentateuch-rest pilot)
• Prior PRs in path: feat(lenses): Exodus 1-10 lens content, batch 1 of Exodus pilot (#820, #1781) #1797 / feat(lenses): Exodus 11-20 lens content, batch 2 of Exodus pilot (#820, #1781) #1798 / feat(lenses): Exodus 21-30 lens content, batch 3 of Exodus pilot (#820, #1781) #1799 / feat(lenses): Exodus 31-40 lens content, batch 4 of Exodus pilot — closes Exodus (#820, #1781) #1800 (Exodus, all merged), feat(lenses): Leviticus 1-10 lens content, batch 1 of Pentateuch-rest pilot (#820, #1782) #1801 (Leviticus 1–10, merged)

[github-actions]

Content Pipeline Results

✅ All pipeline checks passed

Step	Status	Details
Schema Validation	✅	145332 passed, 0 failed
Build DB	✅	—
DB Integrity	✅	101 passed, 0 failed

[github-actions]

Test Results

✅ All tests passed

	Passed	Failed	Total
Tests	✅ 4001	❌ 0	4001
Suites	✅ 532	❌ 0	532

Coverage

Statements	Branches	Functions	Lines
—	—	—	—

⏱️ Duration: 75.6s