harness:status

To monitor LLM agent optimization, you can visualize iteration history and apply heuristics to check if the last three scores are within 1% or if the current best drops below a previous best. This surfaces actionable warnings for regressions or stagnation.

The best way to visualize LLM evolution loop scores is by rendering a score timeline and current metrics. This provides a single-command health check of your optimization run by resolving the evolver tools path and invoking the charting script.

You do not need LangSmith experiment links to check model optimization loop status, but they are used optionally. When available, the system resolves fallback configurations and prints the LangSmith URL for the best experiment.

You can detect LLM regression automatically without manual metric comparison. The system checks whether the current best version is lower than a previous best and surfaces actionable warnings during the evolution loop health check.

To render an evolution chart, your LLM agent project requires access to the .evolver configuration and the evolution_chart.py tool. The system resolves these paths to display the score timeline and current metrics.