← All projects
AI & developer tools
Sentinel
Policy decisions, tamper-evident audit records and replay for consequential application and AI-agent actions.
Governance platformThe idea behind the work
A log alone cannot explain whether an action should have happened. Sentinel brings policy evaluation, evidence retention and replay into a shared service for applications and agent integrations.
Inside the project
- A Go service and CLI provide a central integration point for governed actions.
- OPA policies and chained records connect an action with its decision context.
- Agent integration, operational inspection and archive tooling support later review.
Current scope
An implemented governance codebase. Deployment-specific enforcement and independent assurance require their own verification.