eval:run
eval:run
Section titled “eval:run”Run a certification scenario suite against an agent evaluation report
Options:
| Flag | Description | Required |
|---|---|---|
--json | Output the verdict as JSON | No |
--evidence-dir <dir> | Override the root directory evidence is written under (testing) | No |
--suite <path> | Path to a scenario suite JSON file (defaults to the built-in suite) | No |
--report <path> | Path to the agent evaluation report JSON | No |
--agent-version <version> | Agent version (opaque label) | No |
--agent-id <id> | Agent identifier (opaque label) | No |