pii-masking-patterns

SKILL.md

Protect sensitive data in LLM observability pipelines with automated PII detection and redaction.

| Detection engine | Presidio (enterprise), regex (simple), LLM Guard (LLM pipelines) | | Masking strategy | Replace with type tokens [REDACTEDEMAIL] for debuggability | | Performance | Use async/batch processing for high-throughput | | Langfuse integration | Use mask= callback at client initialization |

| Reversibility | Use LLM Guard Vault for deanonymization when needed |

PII 檢測和屏蔽以實現 LLM 可觀察性。在記錄提示/回應、使用 Langfuse 進行追蹤或保護生產 LLM 管道中的敏感資料時使用。 來源：yonatangross/orchestkit。

查看原文