Preserve Last CI Job Status in Nx Cache to Aid Failures Without Cache

[patels30]

Currently, when a CI task runs in gihtub without cache and fails, that failure is genuine and expected. However, on the next run, Nx stores the result in the cache, even if the previous run failed. This means subsequent executions may incorrectly show a cached success instead of reflecting the earlier failure, leading to misleading feedback and masking real issues. To mitigate this, can we add an optional setting to Nx’s remote cache to persist and expose last CI job status metadata for every target key. When a task runs without cache (miss or disabled)?