Skyflow's pattern for isolating sensitive values in a data privacy vault so prompts, training sets and RAG content carry tokens instead of raw identifiers, with role-based re-identification for authorized processes. It governs sensitive fields rather than discovering data estate-wide, and requires application changes to adopt.
commercial · generally available · Research snapshot 2026-09-06
Show me a RAG flow where the vault de-identifies customer records before embedding and only a privileged role can re-identify the values in the answer.
Capabilities and evidence
Support labels reflect the supplied research. Documentation and vendor claims are not independent product tests. “Not established” means the researcher did not find support; it does not prove a capability is absent.
Documented by provider
A Detect vault stores detected entities and their tokens so data can be de-identified at collection and re-identified under governance controls during training and inference.
Limit: Documentation page describes the pattern, not detection accuracy or entity coverage.
Roles and access controls govern re-identification, with permissions allowing data to be fully or partially revealed only to validated users and authorized processes.
Limit: Does not document audit logging of re-identification events on this page.
Skyflow for GenAI is presented as detecting and redacting sensitive data during collection, model training, fine-tuning, RAG and inference, including data flowing into and out of generative AI agents.
Limit: Product page; agent-specific integration mechanics are not described.