Problem. Agentic systems ship with unknown vulnerabilities and no ongoing
watch; owners learn of holes only after an incident.
Target user. Anyone running agentic apps/services in production.
Market gap. Existing scanners target models or classic apps; continuous
vulnerability scanning + auto-patch for agentic systems is unserved
(Scout research, 2026-05-16).
MVP scope (shipped). Point at agents/apps/services → scan → find
vulnerabilities → apply patch → plain-English report; self-improves per scan.
Acceptance criteria. (1) A scan runs and lists findings. (2) Patch-applied
state is visible. (3) A readable report is produced. (4) Mobile-clean.
Out of scope (v1). Real CVE feeds; live patching of external systems.
Next. Phase 2 — real scan engine + findings store + report generator.