Runtime policy checks for AI agent tool calls
Kontext Runtime hooks tool calls from supported coding agents before execution, scoring actions with deterministic policy and local risk logic, then allowing, asking, narrowing, or blocking them while recording auditable traces.