Phinq is an open-source runtime governance layer positioned between agents and tools. It classifies each call, applies team policy, pauses high-risk or irreversible actions for human approval, and writes decisions to a hash-chained audit log for review and accountability.
Features
- Tool-call risk classification
- Human approval for risky actions
- Policy enforcement
- Hash-chained audit logs
Use Cases
- Constrain production agents
- Protect destructive operations
- Meet audit requirements
- Investigate agent incidents
FAQ
Phinq is an open-source runtime governance layer positioned between agents and tools. It classifies each call, applies team policy, pauses high-risk or irreversible actions for human approval, and writes decisions to a hash-chained audit log for review and accountability. Core capabilities include: Tool-call risk classification, Human approval for risky actions, Policy enforcement.
Common scenarios include: Constrain production agents, Protect destructive operations, Meet audit requirements.
