Running Analysis

Running Analysis

The public build runs Agent Architecture and OpenClaw Architecture metrics. You can run with defaults or use a config file to customize metrics and policy.

Default run

With no config file, Arxo runs the published metrics:

arxo analyze

With a config file

Create arxo.yaml (or .arxo.yaml) to specify which metrics to run and policy thresholds:

metrics:
  - id: agent_architecture
    enabled: true
  - id: openclaw_architecture
    enabled: true

policy:
  invariants:
    - metric: agent_architecture.overall_agent_health
      op: ">="
      value: 70
    - metric: openclaw_architecture.overall_openclaw_health
      op: ">="
      value: 60

Run with:

arxo analyze --config arxo.yaml

Single metric

To run only one metric:

arxo analyze --metric agent_architecture
# or
arxo analyze --metric openclaw_architecture

CI and fail-fast

arxo analyze --fail-fast

Exits with a non-zero code on the first policy violation.

Next Steps

• Configuration — Full config reference
• Policy Examples — Policy patterns
• Agent Architecture — Metric reference
• OpenClaw Architecture — OpenClaw metric reference
• OpenClaw Scoring and Keys — Composite scoring and config semantics
• OpenClaw Policy and CI Gates — Enforce OpenClaw baselines in CI