Researcher Hacks Claude Code Auto Mode via Indirect Prompt Injection from a Website

Security researcher @wunderwuzzi23 demonstrated an attack chain that fully compromises Claude Code Opus 5 auto mode via indirect prompt injection from a website and module shadowing, achieving 60-80% success rates and underscoring the need for mandatory security invariants.

2026-08-30 ~ 2026-09-01 · 4 related posts

• Researcher Breaks Claude Code Opus 5 Auto Mode, Achieving 80% Attack Success Rate — Bedrovelsen · 2026-08-30
• Module Shadowing Attack Hijacks Claude Code for RCE — wunderwuzzi23 · 2026-08-31
• Breaking Claude Code: Hijacking System via Website — wunderwuzzi23 · 2026-09-01
• Claude Code Auto Mode Compromised: Website Summary Leads to RCE — wunderwuzzi23 · 2026-09-01