agent-install

Pass

Audited by Gen Agent Trust Hub on May 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: Credential Protection
  • The instructions explicitly prevent sensitive API keys from being shared in chat sessions or hardcoded in manifests, mandating the use of local environment files and Kubernetes secrets.
  • [EXTERNAL_DOWNLOADS]: The skill downloads the Datadog Operator via Helm and the pup utility via Homebrew from official vendor-controlled repositories.
  • [SAFE]: Indirect Prompt Injection Surface
  • The skill monitors pod logs for authentication errors. This ingestion point is limited to status verification and is followed by a validation check against the official Datadog API. Ingestion points: kubectl logs in Step 5; Boundary markers: None; Capability inventory: kubectl, helm, curl, pup; Sanitization: Grepping for specific authentication error strings.
Audit Metadata
Risk Level
SAFE
Analyzed
May 1, 2026, 05:33 PM
Security Audit — agent-trust-hub — agent-install