RootSign
A library that logs the actions AI agents take in production as a tamper-evident, cryptographically hash-chained record, so teams can prove what an agent did and detect if a log was altered afterward.
RootSign captures each action an AI agent takes in production — calling tools, hitting APIs, writing to databases — as a record containing a SHA-256 hash of the previous action, forming a hash chain. The rootsign verify command detects any after-the-fact tampering with the log.
Installation is a single pip install rootsign command with no Docker or database required. It integrates with LangGraph and CrewAI, includes a framework-agnostic MCP proxy, PII redaction, human-in-the-loop checkpoints, and can export evidence bundles for compliance use.
It is written in Python and is at v0.3.0, with a hosted ingest backend and compliance dashboard planned.