langsmith-observability
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references the langsmith library, which is a well-known and trusted library for LLM observability.
- [INDIRECT_PROMPT_INJECTION]: The skill demonstrates patterns for processing external datasets and model traces. Ingestion points: Data is ingested via evaluate() from datasets and pull_prompt() from the LangSmith Hub in SKILL.md and advanced-usage.md. Boundary markers: Not explicitly defined in the provided SDK usage patterns. Capability inventory: The skill utilizes network operations to communicate with the LangSmith API and wraps model functions for tracing. Sanitization: The skill explicitly includes a sanitize_inputs example to redact sensitive fields like passwords before transmission.
Audit Metadata