What it does
A proof-of-behavior layer for AI agents. You write permit, forbid and require rules, enforce them at runtime, and keep hash-chained audit logs that reveal later tampering. The repository describes itself as an earlier direction, while the newer prototype lives in a separate repository.
Use cases
- 01Write rules that forbid certain agent actions
- 02Require a check before a sensitive action runs
- 03Keep tamper-evident logs of agent decisions