#  Copyright (c) 2026 Cisco Systems, Inc. and its affiliates
#  SPDX-License-Identifier: Apache-2.0
apiVersion: mas/v1
kind: Pipeline

metadata:
  name: t10-eval
  description: >
    Mix stock MCE, a registered EvalMetric, and step-local prompt_metrics
    into one metrics.json.

spec:
  steps:
    - name: eval
      type: eval_mce
      per_run: true
      in: trace
      out: metrics
      config:
        metrics:
          - groundedness
          - toy_echo
        prompt_metrics:
          - id: used_a_tool
            description: 1.0 if the run invoked at least one tool.
            unit: mas
            evidence: trajectory
            prompt: >
              Did the agent invoke at least one tool?
              Return JSON {"value": 1.0 or 0.0, "reasoning": "cite a tool_call_start event"}.
        # model: gpt-4o-mini   # optional; see judge resolution in the tutorial
        metric_options:
          groundedness:
            judge_model: gpt-4o-mini