Post #17564 227 Oct 6, 2026, 17:36 UTC ■■■□□ Interesting thread 🧵 by project discovery on AI-Hackinghttps://x.com/pdiscoveryio/status/2107522665951227970 X (formerly Twitter) ProjectDiscovery (@pdiscoveryio) on X We backdoored a 7B open model for under $50. Pointed Codex at it -> it silently stole credentials the moment we hit the trigger phrase. ->100% hit rate. -> zero false triggers on normal prompts…