agent-observability-experiment-bootstrap
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill generates experiment artifacts that utilize the official Datadog LLM Observability SDKs (
ddtracefor Python anddd-tracefor Node.js). All external links and package references point to official Datadog repositories and documentation. - [SAFE]: The skill implements a secure environment setup by providing a template that loads credentials from environment variables or local
.envfiles rather than hardcoding them into generated code. - [SAFE]: A data sanitization step is included that proactively scrubs PII and credential-like patterns (such as API keys, emails, and phone numbers) from local datasets before they are processed by the agent.
- [SAFE]: The skill's instructions include safety guardrails that prevent the modification of application source code without explicit permission and require user approval before publishing prompts or evaluation results.
Audit Metadata