Function shell_init_errors 

pub fn shell_init_errors(
    ctx: &ExecutionContext,
    runs: usize,
) -> Result<ProbeResult>

Render the cross-history errors view.

Scans the last runs profiles, keeps only entries with non-zero exit status, groups them by target, and orders by failure count (most-broken first). The runs parameter follows the same window convention as shell_init_filter.