Last released Oct 8, 2026
A LangChain middleware that checks each tool call of an agent against the permission a person signed, and records it with Provared. Evidence, not a verdict.
Proof of what an AI agent was allowed to do, and what it then did. Evidence, not a verdict.