argocd-aks-cluster-registration
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external command outputs (e.g., Terraform JSON) which is then used as arguments in shell commands. Ingestion points: The REG variable in Workflows/RegisterCluster.md is populated by an external command output. Boundary markers: No specific delimiters are used when interpolating the extracted values into the argocd command. Capability inventory: The skill executes az, kubectl, and argocd commands. Sanitization: The skill uses jq to validate and extract specific fields (client_id, tenant_id) from the input, which mitigates simple malformed data but relies on the integrity of the source command.
- [COMMAND_EXECUTION]: The skill makes extensive use of system CLI tools to manage infrastructure. Evidence: Use of az aks get-credentials, argocd cluster add, and kubectl exec to perform environment probes and cluster registration. Evidence: A curl command is used to send notifications to a local service at http://localhost:31337/notify.
Audit Metadata