Data Failover Skill

To keep security scans running when APIs fail, you need an API failover chain that automatically switches to backup providers or marks data as unavailable. This ensures scan execution continues without stalling during partial outages.

LLM fallback sequencing is a prioritized model sequence for main and sub-agents that maintains progress when specific model endpoints fail. It applies fallback ordering to ensure pipeline resilience during intelligence scans.

To handle token scoring with unavailable data sources, score with whatever data is available and annotate the missing data. Never block a scan due to source downtime; retry fetching the missing data on the next cycle.

Pipeline resilience logic works for bounty research and contract verification by applying per-source fallback ordering across dex, wallet, social, and contract tasks. It maintains autonomous scan continuity during expected partial outages.

When many intelligence APIs are down during a scan, failover logging and alerts are triggered. The system scores with available data, marks missing sources as skipped, and retries on the next cycle without blocking execution.