OpenClaw - The tweet raises concerns about prompt injection risks

The tweet raises concerns about prompt injection risks when using an agent's email address, and mentions using OpenClaw as part of a 'paperclip setup'.

Updated: 4/16/2026
@bensrmn @dotta Doesn’t this unnecessarily expose it to prompt injection risks? If your agent’s email address is known to others, malicious people could email to that address with prompt injection attacks. In my paperclip setup, I added OpenClaw as an agent, and this allows it to help me Source: https://x.com/RoboAlchemist/status/2038817891168965097

Did this solve your problem?

0 developers found this helpful