eks-recon
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted metadata from the EKS environment, such as pod labels, container names, and service configurations, which could potentially contain adversarial instructions designed to influence the agent's behavior during the reporting phase.
- Ingestion points: Data is collected through numerous discovery commands across all sub-agents, including
kubectl get pods -A,kubectl get services -A, andaws eks describe-cluster. - Boundary markers: The instructions do not define specific delimiters or "ignore previous instructions" guards for the data retrieved from the Kubernetes and AWS APIs.
- Capability inventory: The skill utilizes powerful CLI tools such as
aws,kubectl, andhelmto perform its discovery mission, providing a broad surface for data collection. - Sanitization: There is no evidence of sanitization or filtering applied to names, labels, or annotations retrieved from the cluster before they are incorporated into the final Markdown and YAML reports.
Audit Metadata