Module policy_engine 

Structs

Verdict

Enums

VerdictStatus

Functions

evaluate_schema

Evaluates tool arguments against a compiled schema.

evaluate_sequence

Evaluates a sequence of tool calls against a sequence policy (regex-like). For v0.9, simplified: the policy is just a string (regex) of tool names. E.g. “^search (analyze )*report$” The input is a list of tool names invoked in order.

evaluate_tool_args

Evaluates tool arguments against a policy (JSON/YAML Value). The policy is expected to be a map of tool_name -> schema.