# Webdev Agent 100-Point Audit Scorecard

The audit uses weighted categories. Each category receives a score from 0 to its maximum weight.

## Categories

| Category | Weight | What is evaluated |
|---|---:|---|
| Strategy and clarity | 10 | Audience, offer, purpose, CTA, success definition |
| Content and trust | 10 | Accuracy, credibility, proof, objections, contact and policy information |
| Visual design | 12 | Hierarchy, typography, spacing, consistency, brand fit, polish |
| UX and conversion | 15 | Navigation, friction, forms, CTA placement, feedback, task completion |
| Responsive quality | 10 | Mobile, tablet, desktop behavior; touch targets; overflow |
| Accessibility | 10 | Semantics, keyboard, focus, contrast, labels, alternatives |
| Performance | 10 | Loading, assets, rendering, caching, responsiveness |
| Technical reliability | 10 | Errors, state handling, validation, browser behavior, integrations |
| Security and privacy | 8 | Secrets, access control, input handling, TLS, privacy behavior |
| SEO and discoverability | 3 | Titles, descriptions, indexing, structured content, sharing metadata |
| Operations and maintainability | 2 | Documentation, logging, deployment, rollback, ownership |

**Total: 100 points**

## Scoring guidance

### 90–100: Premium / production ready

The experience is distinctive, polished, reliable, measurable, and ready for public use. Remaining issues are minor refinements.

### 80–89: Production ready with minor improvements

The project is safe and useful for production. There are visible opportunities to improve polish, performance, content, or maintainability.

### 70–79: Conditional release

The project may function, but important quality gaps remain. Release only with explicit acceptance and a near-term remediation plan.

### 60–69: Prototype only

The project demonstrates the concept but lacks the reliability, refinement, or operational readiness expected of production software.

### Below 60: Major remediation or rebuild

The experience has structural weaknesses that cannot be solved by small cosmetic changes alone.

## Audit method

1. Test the primary user journey first.
2. Review at mobile, tablet, and desktop widths.
3. Use keyboard-only navigation.
4. Trigger validation and failure states.
5. Inspect network, console, and server errors.
6. Verify real content, links, policies, and contact routes.
7. Check deployment, logging, backup, and rollback documentation.
8. Record evidence and recommended action for every deduction.
9. Mark release blockers separately from the numerical score.

## Priority system

- **P0 — Release blocker:** security, privacy, broken critical flow, data-loss risk
- **P1 — High:** major conversion, reliability, accessibility, or mobile failure
- **P2 — Medium:** noticeable quality or maintainability problem
- **P3 — Low:** polish, consistency, or optimization opportunity

## Recommended audit report fields

- Project and version
- Audit date and auditor
- Environment tested
- Primary flow tested
- Category scores
- Total and classification
- Release blockers
- Findings with priority, evidence, recommendation, and owner
- Retest date