coreweave-multi-env-setup
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill demonstrates high security maturity by explicitly recommending secrets managers over local environment files and using clear placeholders for sensitive data like service account tokens.
- [COMMAND_EXECUTION]: Bash usage is strictly scoped to
kubectl,kustomize, andcurlfor benchmarking. The YAML frontmatterallowed-toolsfield correctly implements the principle of least privilege by restricting bash to specific binaries. - [INDIRECT_PROMPT_INJECTION]: The skill reads status information from the Kubernetes API via
kubectl. While this constitutes an ingestion point for external data, the implementation uses specificjsonpathfilters to limit the data returned to the agent, which is a standard and safe operational pattern for infrastructure management skills. - [DATA_EXPOSURE_EXFILTRATION]: Network activity is limited to benchmarks against the CoreWeave cloud service (coreweave.cloud) and standard Kubernetes API interactions. No patterns of credential harvesting or exfiltration to unauthorized domains were detected.
Audit Metadata