The Agent Skills Directory

[COMMAND_EXECUTION]: The skill explicitly instructs the agent to execute user-provided integration tests. Step 1 states: 'If a user selects a testcontainer test, please run it locally and verify that it passes.' Since these tests are external code provided at runtime, this creates a significant risk of arbitrary code execution on the agent's host system.
[COMMAND_EXECUTION]: The skill uses Terraform's local-exec provisioner to execute helmfile apply. While a standard infrastructure-as-code pattern, it represents a capability to execute shell commands from generated files.
[EXTERNAL_DOWNLOADS]: The skill references Helm charts hosted at https://voltdb-kubernetes-charts.storage.googleapis.com. This is identified as a vendor-owned resource for VoltDB and is used for its primary purpose of cluster deployment.
[PROMPT_INJECTION]: The skill instructs the agent to 'verify with a user that there is a running integration test' and 'quit' if one is not provided. While intended as a safety check for deployment validity, the logic depends on unvalidated user input and external code execution.
[DATA_EXFILTRATION]: The skill manages sensitive credentials, including Docker registry authentication and VoltDB licenses. While it correctly suggests using .gitignore for *.tfvars and *.local files, the presence of these credentials in the local environment where unverified 'integration tests' are run increases the risk of credential exposure.

volt-kubernetes