Re:Claude Cowork Exfiltrates Files
by niyikiza on Wednesday January 14, 2026 @05:28PM
(#46624698)
Exactly. I'm experimenting with a "Prepared Statement" pattern for Agents to solve this:
Before any tool call, the agent needs to show a signed "warrant" (given at delegation time) that explicitly defines its tool & argument capabilities.
Even if prompt injection tricks the agent into wanting to run a command, the exploit fails because the agent is mechanically blocked from executing it.
Before any tool call, the agent needs to show a signed "warrant" (given at delegation time) that explicitly defines its tool & argument capabilities.
Even if prompt injection tricks the agent into wanting to run a command, the exploit fails because the agent is mechanically blocked from executing it.