get-results

To retrieve simulation results from the Coval evaluation platform, use this skill to fetch and parse data by providing a specific run or simulation identifier. It extracts status, transcripts, and error diagnostics for performance analysis and debugging.

Debugging failed AI evaluation simulations is best handled by parsing run outcomes to identify failed test cases. This skill automatically extracts error messages and transcripts, allowing you to focus immediately on debugging specific failures from the evaluation run.

You can generate a summary table of passed and failed simulations after an evaluation run by pulling detailed simulation logs. The skill parses status and error diagnostics to summarize outcomes, facilitating quick performance analysis and targeted debugging.

The get-results skill requires unique run or simulation identifiers to fetch evaluation outcomes. It operates on these specific identifiers to programmatically extract status logs, transcripts, and error diagnostics via the Coval CLI.

Missing error diagnostics in AI simulation transcripts may occur if the correct run identifier is not used during retrieval. This skill parses simulation data to explicitly extract and surface error messages, ensuring failed test cases are identified.