ai-inference-service-mesh
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No patterns of instruction override, jailbreaking, or safety bypass were detected in the skill content.\n- [DATA_EXFILTRATION]: No sensitive data access or unauthorized exfiltration attempts were found. The use of ServiceEntry for api.openai.com is a standard configuration for managing legitimate external AI service traffic.\n- [COMMAND_EXECUTION]: The skill utilizes standard Kubernetes (kubectl) and Istio (istioctl) CLI tools for their intended infrastructure management and observability purposes.\n- [PRIVILEGE_ESCALATION]: While the skill documents operations requiring cluster-administrative access (e.g., mesh installation and namespace labeling), these are expected and necessary for the stated DevOps purpose.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies potential attack surfaces such as routing based on client headers and provides explicit defensive guidance, such as validating and stripping headers at trusted gateways, to mitigate risks.
Audit Metadata