Final Pre-Publication Audit Report

Audit date: 2026-09-18 Scope: entire repository, judged on executed checks rather than source inspection alone Overall status: PASS

Every result in this report comes from a command actually executed during this audit session. Where a check produced a finding, the finding is listed under "Publication blockers" or "Recommended corrections" with its evidence; no result is claimed without execution.

1 · Curriculum — PASS

Check	Result
All 32 lectures exist	✓ `lectures/lecture-01` … `lecture-32`, each with `plan.md`, `notes.md`, `teaching-notes.md`
Foundational → advanced progression	✓ Block I (L01–08) problem formulation → Block II (L09–16) data/strings → Block III (L17–24) algorithms → Block IV (L25–32) graphs, simulation, optimisation, capstones
Learning outcomes documented	✓ `docs/learning-outcomes.md`; LO mappings present in all 32 plans
Two-hour plan realism	✓ Every plan carries a 120-minute table with a break and a flex column; workshops and reviews use the documented alternate structure
Case distribution across lectures	✓ 144 cases over 32 lectures (120 base at exactly 4/lecture + 16 applied additions, each mapped by metadata)
Assessment & lab integration	✓ 12 labs, 8 quizzes, 6 assignments, 3 projects, 5 activities — each mapped to lectures and LOs

Tier distribution of the 144 cases: 33 Beginner / 37 Foundational / 38 Intermediate / 23 Advanced / 13 Expert — every requested minimum (20/25/25/20/10) exceeded.

2 · Case studies — PASS

Check	Result
≥ 100 unique cases	✓ 144 (cs-001 – cs-144), IDs unique, no gaps
Clear IDs and difficulty levels	✓ every case carries id, difficulty, lecture, topics, estimated time, prerequisites, objective in its metadata table
Complete student statements	✓ all five required student sections present (validator-enforced)
Complete instructor solutions	✓ all twelve required instructor sections present (validator-enforced)
Correct solution mapping	✓ 144/144 one-to-one pairs, verified by `scripts/validate.py`
Prerequisites stated	✓ per-case prerequisites reference specific lectures
Edge cases and assumptions	✓ required sections in every solution
Verified expected outputs	✓ every reference program executed twice (determinism confirmed); expected-result lines rewritten from actual output during authoring
No unintentional duplicates	✓ duplicate-title scan over case metadata and deck titles: 0 duplicates

Prior audits on this corpus: docs/case-review.md (128-case pedagogical review), docs/instructor-reports/expert-tier-rationale.md (all 12 base Expert cases), docs/instructor-reports/quiz-answer-key-audit.md, docs/instructor-reports/lab-expected-output-audit.md.

3 · Teaching material — PASS

Student notes: plain-language, pseudocode before Python throughout.
Instructor notes: board plans, question scripts, misconception lists in all 32.
Python examples: solution-execution harness runs every case's reference program; all 144 green, deterministic across repeated runs.
Pseudocode: house style enforced by docs/validator conventions.
Scaffolding: Expert cases carry the rationale and prerequisite chain in docs/instructor-reports/expert-tier-rationale.md.

4 · Security and content separation — PASS

Checked the built public artifact (site/), not just source names:

Check	Result
Instructor case files absent from public output	✓ zero `cs-*.md`/HTML beyond the 144 public student pages
Leak-phrase scan of built site	✓ 9 hits, all policy descriptions on handbook pages (contributing/deployment/faq) — no leaked content
`def` /`print(` on case pages	✓ matches are prose in student debugging bullets, not code blocks; student sources contain zero Python fences
Instructor-only reports (`docs/instructor-reports/`)	✓ excluded from the public build by name; present in the separate `public_instructor/` instructor build (572 pages, gitignored)
Secrets / API keys / credentials	✓ grep for key/password/secret patterns: only match is the legitimate case title "Password Rules" — no credentials
Workflow artifact source	✓ `pages.yml` uploads `site/` only; `build_site.py` fails the build on any leak, so instructor material is physically excluded

5 · Website — PASS

Check	Result
Build succeeds	✓ `python scripts/build_site.py` completes; leak check passes
Internal links	✓ site audit: 367 pages, 9,691 references, 0 broken
Navigation	✓ handbook nav includes all lectures, case catalog, labs, quizzes, assignments, projects, heatmap; instructor-only pages absent from student nav
Mobile & desktop	✓ responsive CSS with viewport meta present on built pages
Assets	✓ CSS resolves; heatmap SVG is well-formed XML (parser-verified); CSV links resolve
Structure	✓ single `h1` per page, hierarchical headings enforced by audit

6 · GitHub Actions — PASS

Check	Result
Correct location	✓ `.github/workflows/` at repository root; no nested repository
Coherent build & validation	✓ `pages.yml`: validate → heatmap → slides → build → leak check → upload → deploy; `validate.yml`: validation only, no publish steps
Pages architecture	✓ modern `actions/upload-pages-artifact` + `actions/deploy-pages` flow; requires the documented Settings → Pages → Source: GitHub Actions choice
No competing deployment workflow	✓ exactly one deployment workflow
Maintained action versions	✓ `actions/checkout@v4`, `actions/setup-python@v5`, `actions/configure-pages@v5`, `actions/upload-pages-artifact@v3`, `actions/deploy-pages@v4`
Gate chain	✓ each stage blocks the next on failure; validation failure blocks deployment
No commits or pushes	✓ none performed; workflows trigger only on the user's own push

7 · Documentation — PASS

Check	Result
README complete	✓ overview, structure, setup, preview, validation, deployment, license
Local setup works	✓ documented commands resolve to real scripts
Local preview documented	✓ `python -m http.server` instructions for `site/` present
Validation commands documented	✓ `python scripts/validate.py`, `python -m pytest`, `python scripts/qa_report.py`
Pages deployment documented	✓ `docs/deployment.md` covers the Settings choice, workflow stages, instructor exclusion
Instructor distribution documented	✓ `docs/instructor-guide.md` and `build_instructor.py` produce a separate gitignored instructor build
License & attribution	✓ dual license: MIT for tooling, CC BY 4.0 for content; no bundled third-party datasets or invented external references

Checks performed (executed)

python scripts/qa_report.py — full 7-stage suite → PASS
python scripts/validate.py — 144/144 cases, 0 errors, 0 warnings
python -m pytest tests/ -q — 28/28 passed
Site build + python scripts/audit_site.py — 367 pages, 0 broken links
Instructor-build dry run — 572 instructor pages to a gitignored directory
Leak scans (instructor filenames, leak phrases, code fences) on site/
Secrets grep across the repository
Duplicate-title scan across cases and decks
Tier/lecture distribution computed from case metadata
Workflow chain inspection (all gates present, single deploy workflow)
Heatmap regeneration (23 topics × 32 lectures over 144 cases)

Publication blockers

None.

Recommended corrections (non-blocking)

First deployment is manual-once: Settings → Pages → Build and deployment → Source: GitHub Actions. The workflow cannot set this itself; everything after it is automatic.
Instructor distribution: instructors use python scripts/build_instructor.py and share public_instructor/ out-of-band (the directory is gitignored and never deployed). Document any local sharing policy your department requires.
Two cosmetic warnings the QA suite reports as informational: 14 lectures carry more than 4 cases (intentional applied additions cs-129–cs-144) and those additions include no Beginner-tier case above cs-142's lecture — both are by design and listed as information, not defects.

Conclusion

The repository is publication-ready: 144 verified cases across 32 coherent lectures, complete teaching and assessment material, a leak-free public site build with zero broken links, a single coherent deployment workflow, and documentation covering setup, preview, validation, deployment, and instructor separation. No commits or pushes were performed by this audit.