Skip to main content
agentcompass summary recomputes Benchmark aggregate metrics from task results in an existing run directory and updates the metric files and their provenance. It does not rerun agents, verifiers, or analyzers.

Regenerate a Summary

RUN-DIR should point to the complete AgentCompass result directory, not its details/ subdirectory:
AgentCompass strictly validates saved task details and the attempt plan in run_info.json, then rebuilds one Metric Contract report. By default, it replaces summary.md and metrics.json in place without modifying task results, and records source: "summary" plus the report plan under run_info.json.metric_artifacts.

Preview the Summary

Use --dry-run to preview the new summary.md in the terminal before writing files:
This mode does not update any files in the run directory.

Parameters

You normally do not need --config or --benchmark-params. Use them only to supply or adjust current Benchmark aggregation settings such as aggregation_mode or category_hierarchy. They cannot change the persisted execution.attempts plan and do not reevaluate tasks. In a non-dry-run regeneration, metric_artifacts.benchmark_params_override stores the redacted --benchmark-params object; the original request and params.json remain unchanged.