🧮 The TOVCDI Code Review Rubric
Your third always-on Featured resource. A shared standard for reviewing code — your own or a peer's — scored against the B.U.I.L.D. Method. Use it in "Audit Anything" posts and in the Capstone.
How to use it
Score each of the 5 dimensions 0–2. Add them up (max 10). The goal isn't a perfect 10 every time — it's knowing exactly where the work stands and what to fix next.
| Score | Meaning |
|---|---|
| 0 | Missing / not addressed |
| 1 | Started, but gaps remain |
| 2 | Solid — a developer would sign off |
The 5 dimensions
1. ✅ Correctness (the B) — does it actually work?
- Happy path works and edge cases are handled (empty, huge, weird, double inputs)
- No console errors; no crash on bad input
- 2 = "I tried to break it and couldn't." 0 = "works if you do exactly the right thing."
2. 🧠 Clarity (the U) — can a human understand it?
- Names say what things are; functions do one thing
- No dead code, no mystery blocks the author can't explain
- 2 = a stranger could read it. 0 = even the author isn't sure what a part does.
3. 🎨 Experience (the I) — is it usable by real people?
- Custom (not default-AI) look; responsive on a phone
- Loading/empty/error states exist; errors are human
- ♿ Accessible: semantic tags, alt text, keyboard-usable, contrast
- 2 = pleasant on any device, for any user. 0 = desktop-only, divs-as-buttons.
4. 🔒 Security (the L) — is it safe?
- No secrets in code; input validated server-side; auth enforced on the backend
- Database/permissions locked down; dependencies audited
- 2 = passes the Security Audit Checklist. 0 = keys in code or an open database.
5. 📦 Ship-readiness (the D) — will it survive reality?
- README, version control with real commits, the critical path tested
- Deploys cleanly; env vars set in prod; a rollback plan exists
- 2 = someone else could run and deploy it. 0 = "works on my laptop only."
🟢 Score guide
- 9–10 — Ship it. This is developer-grade work.
- 6–8 — Almost. Knock out the 1s and 0s, then ship.
- 3–5 — A solid draft that needs the back half of B.U.I.L.D. Great learning target.
- 0–2 — A vibe-coded start. Perfect — now you know the whole roadmap ahead.
🤝 How we give feedback here (the culture)
- Start with what works. Name one genuinely good thing first.
- Critique the code, encourage the coder. "This function does X; it'll break on Y" — not "this is bad."
- Be specific and actionable. Point to the line, suggest the fix, link the relevant lesson.
- Ask before assuming. "What was your thinking here?" beats "why did you do this?"
A review is a gift, not a verdict. We review to level each other up.
Build It Right, Or Don't Build It At All. 🏛️