otel-genai-instrumentation
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides documentation and code examples for observability instrumentation using OpenTelemetry, specifically for the Honeycomb platform. All code samples are standard implementations of the OpenTelemetry API and SDK.
- [SAFE]: All suggested dependencies and remote references target official OpenTelemetry repositories or standard AI provider SDKs. This includes packages like
opentelemetry-instrumentation-openai-v2and@opentelemetry/instrumentation-openai. - [SAFE]: The skill includes a "Critical Requirements" section that mandates the agent to ask for user consent before enabling the capture of potentially sensitive data (prompts and model responses). It provides clear warnings about PII and proprietary data.
- [SAFE]: Extensive guidance is provided for privacy controls, including message filtering, data truncation, and environment-specific capture strategies (e.g., full capture in dev, redacted in prod).
- [SAFE]: No malicious patterns such as prompt injection, data exfiltration to unknown domains, or obfuscated commands were detected.
Audit Metadata