langchain-agent-monitoring
Pass
Audited by Gen Agent Trust Hub on Nov 20, 2026
Risk Level: SAFE
Full Analysis
- Operational Guidance: The skill provides detailed instructions for using the
langsmithCLI and API to monitor production agents. This includes managing evaluators, rules, and insights jobs, which are standard features of the LangSmith platform. - Input Validation Considerations: The instructions include proactive security advice for building LLM-based judges, specifically recommending that trace content be treated as 'untrusted evidence' rather than instructions. This is a best practice for preventing indirect prompt injection from influencing evaluation results.
- Authentication Best Practices: The skill correctly advises against using
langsmith auth tokenand instead points to the official authentication mechanisms, promoting secure handling of platform credentials. - Official Resource References: All external documentation and API references point to the vendor's official domains (
langchain.com), and the recommended tools (such as thelangsmithCLI andjq) are standard utilities within the developer ecosystem.
Audit Metadata