Repository navigation
Phase 4: AI Reporting & SMB-Friendly UX #3
Description
Phase 4 — AI Reporting & SMB-Friendly UX
Depends on: Phase 3
Drips Wave alignment
- Maps to Wave #29 secure key storage — encrypted
ANTHROPIC_API_KEYat rest - Plain-English output for non-specialist users (dual-audience design in PRD)
Deliverables
- Grounded AI reports with finding ID citations in
ai_reporter.py - Per-OWASP-category remediation cards in dashboard
- Documented safety score algorithm (A–F grading)
- Optional passphrase-encrypted API key storage module
- Template fallback quality audit (no hallucinated CVEs)
- Presentation slides 9–10 screenshots from live demo
Acceptance criteria
- AI report cites only findings present in scan results
- Dashboard shows plain-English "how to fix" per finding
secrets.validate_settings()warns on production without encryption- Demo workflow completable in < 5 minutes for viva panel
References
- PRD §4 Solution
- FUT Minna thesis Chapter 5
Activity
huangji6693-max commentedon Oct 27, 2026
Submitted PR #8 implementing Phase 4 grounded AI reporting (
[ID: <finding.id>]citations and anti-hallucination CVE filter), per-OWASP-category plain-English remediation cards in the dashboard, documented A–F safety score algorithm, PBKDF2-HMAC-SHA256 passphrase-encrypted API key storage (enc:v1:...) with production validation warnings, and viva demo documentation: #8