Helm local index cache interaction with CI/CD chart deployment

The goal is to guarantee that Helm selects the most recent chart version available in a remote repository when executing install, upgrade, template, dependency update, or lint commands inside an automated CI/CD workflow, without requiring a manual helm repo update step.

However, Helm’s local repository index cache has no built‑in TTL and is reused indefinitely until explicitly refreshed. This can cause the cached index to serve stale chart metadata, leading to older chart versions being pulled or referenced, which affects reproducibility and may introduce known vulnerabilities. The same staleness propagates to any Helm sub‑command that reads the index.

• Can a TTL or automatic refresh be configured for the local index cache without patching Helm?
• What pipeline‑level practices ensure the index is updated before any Helm operation runs?
• How does reliance on a stale index affect version consistency across multiple environments that share the same chart repository?