Elevate DesignOS quality system and onboarding

[ardamoustafa1]

What changed

• added a Brief Compiler that converts vague requests into testable design contracts and derived visual directions
• expanded the Design Loop with direction exploration and rendered viewport inspection
• added Brand Fit & Distinctiveness as a seventh review dimension
• changed the CLI to distinguish static risk indicators from unassessed final design scores and enforce blocker caps
• expanded WCAG 2.2 guidance and evidence-led final-gate requirements
• rebuilt the README below its preserved hero with clearer installation, agent setup, diagrams, CLI guidance, examples, and troubleshooting

Why

The previous system contained strong design knowledge, but common prompt entry points could bypass originality and render inspection. Static checks could also appear more authoritative than their actual coverage. This update makes the process clearer, more honest, and easier for new users to adopt.

Impact

Users get a more structured onboarding path, stronger non-generic design prompts, clearer evidence boundaries, and a review workflow that reports unmeasured dimensions as NOT ASSESSED.

Validation

• npm test — 26 passed
• npm run validate — references, accessibility basics, and token drift clean
• npm run check:links — clean
• git diff --check — clean
• README hero prefix verified byte-identical to the original

[github-actions]

DesignOS Review

Files reviewed: 5
Findings: 0
Static risk floor: 100/100 (not a final design score)
Final design score: NOT ASSESSED

Dimension	Static risk indicator
UI Craft	100
UX & Flow	100
Accessibility	100
Modernity	100
Conversion	100

Not assessed: visual hierarchy and brand fit; keyboard completion; accessibility tree and screen reader; computed contrast across real states; 320px reflow/400% zoom; reduced motion; Lighthouse/Core Web Vitals; task completion and conversion quality.

No deterministic findings. Run the human/model review loop for taste, hierarchy, and fit.

Agent-ready fix prompt

Fix these DesignOS review findings, then rerun the DesignOS loop.
Target: examples
Current static risk floor: 100/100. Requested mechanical floor: 95/100.
Final design score: NOT ASSESSED until the manual/browser evidence ledger is complete.

Rules:
- Do not trust the agent self-score; the final pass is this deterministic review output.
- Do not invent customers, testimonials, certifications, usage counts, awards, urgency, or compliance claims.
- Do not claim SOC2, ISO 27001, HIPAA, PCI, GDPR, FIPS, audited, or certified unless a real source is linked.
- Keep visual direction unless a finding directly contradicts it.

Required fixes:
- No deterministic findings remain. Run the human/model review loop for taste, hierarchy, and fit.

Run locally with node DesignOS/bin/designos.js review <path> --min 95.