Incidents in AI-built software, rated and sourced.

What broke, how it happened, and how to check your own app. Severity is our call; the method is on the methodology page.

Log

Severity	Incident	Date
HIGH	RedAccess scan: 380,000 vibe-coded assets, about 5,000 corporate apps, roughly 2,000 exposing private data Lovable, Replit, Base44, Netlify	2026-05
CRITICAL	Lovable BOLA flaw left older projects readable for 48 days after it was reported Lovable	2026-03-03
CRITICAL	Moltbook exposed about 1.5M agent API keys through an open Supabase database Supabase (no RLS)	2026-01-31
HIGH	CVE-2025-53773: prompt injection could switch GitHub Copilot into auto-approve ('YOLO') mode and run commands GitHub Copilot	2025-08-12
HIGH	Claude Code: several 2025 CVEs let untrusted content bypass command confirmation prompts Claude Code	2025-08-01
MEDIUM	Slopsquatting: AI coding models invented 205,474 unique package names in one study All LLM coding tools	2025-08
HIGH	Replit agent deleted SaaStr's production database during a code freeze Replit Agent	2025-07-18
CRITICAL	Base44 auth bypass let outsiders register for private, SSO-only apps Base44 (Wix)	2025-07-09
MEDIUM	Amazon Q Developer VS Code extension 1.84.0 shipped with an injected 'wipe' prompt Amazon Q Developer	2025-07
CRITICAL	CVE-2025-48757: missing Row Level Security in Lovable-generated apps (170 of 1,645 scanned) Lovable + Supabase	2025-03-20
HIGH	Enrichlead: a SaaS built with 'zero hand-written code' in Cursor was abused within days and shut down Cursor	2025-03-17
HIGH	CurXecute (CVE-2025-54135): prompt injection could make Cursor write and auto-start an MCP server Cursor	2025