We found a RubyGems flaw that could leak API keys. πŸ”Ž Researchers say OpenAI agents tried exploiting it 55 days before we reported it. RubyGems patched it, revoked all legacy keys, and found no evidence the theft succeeded. Read πŸ‘‡

Sep 16, 2026 Β· 6:57 PM UTC

3
1
12
1,061
Link to research: trufflesecurity.com/blog/rog…

We discovered a Ruby account takeover; Rogue OpenAI Agents exploited it 2 months prior β—† Truffle...

We discovered and reported a RubyGems vulnerability in July. Researchers now believe OpenAI agents had attempted to exploit it nearly two months earlier.

trufflesecurity.com
156
Sort replies: Relevant Recent Liked
Replying to @trufflesec
Short-lived keys are what makes a timeline like that harmless: with 30-day rotation the window they probed was already dead.
12
Replying to @trufflesec
Agents hunting keys before humans report them is the new timeline. I still scan whatever they wrote before it hits git. safeweave.dev

SafeWeave β€” The AI Code Security Platform

Secure AI-generated code before it ships. 8 scanners that find vulnerabilities, secrets & risky dependencies inside Cursor, Claude Code, VS Code & Windsurf β€” in seconds.

safeweave.dev
3
Load more