Hijacking multi-agent systems in your PajaMAS
Summary: Trail of Bits released pajaMAS, a set of demos showing 'MAS hijacking,' an attack that manipulates the control flow of multi-agent systems by exploiting how agents communicate, described as a variant of prompt injection targeting MAS control flow and discovered by Triedman et al. The demos cover attack vectors including a malicious environment, a malicious tool, and a malicious agent, plus architectural weaknesses such as the 'lethal trifecta,' cycles, and naive security controls.
What you should know: The source states that adding more agents typically multiplies uncertainty instead of adding robustness, and that agents cannot reliably enforce security boundaries because they can be manipulated to lie about security status - so controls bolted onto a single node of a MAS can be circumvented.
Read the original at Trail of Bits Blog ↗
- Topics
- AI Security
- Related
- SECURITY AFFAIRS AI-CYBERSECURITY NEWSLETTER ROUND 8 Cyber Risks · Nov 15, 2026Pentagon Renames CDAO ‘Super Intelligence Office’ Cyber Risks · Nov 15, 2026186 Fake AI Domains Redirect Users to Click Fix Lures, Browser Malware, and Search Hijackers Cyber Risks · Nov 12, 2026