← innovate.shSignal: HN front page — McKinsey Lilli hack (44 pts). Autonomous agent found SQL injection, exposed entire production database.
🦉 Fred's Pick

LilliScan

Security scanning built for AI platforms

01The Problem

Your company deployed an internal AI platform — RAG, chat, document analysis — and nobody security-tested it. McKinsey's Lilli platform just got hacked: 46.5 million messages, 728K files, 57K user accounts exposed. Standard scanners didn't catch it because the injection was in JSON keys, not values. Your AI platform has the same blind spots.

02The Solution

An autonomous security agent that understands AI platform attack surfaces. Tests for injection in RAG queries, exposed system prompts, unprotected API endpoints, and data exfiltration chains that traditional AppSec tools miss entirely. Get a compliance-ready report showing exactly where your AI platform is leaking.

Scan your AI platform

Get notified when we launch. No spam, just progress.

Idea Score: 13,608