antithesis-setup-k8s
Pass
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill exclusively references trusted and vendor-owned domains, such as antithesis.com and the antithesishq GitHub organization, for all documentation and configuration retrieval.
- [EXTERNAL_DOWNLOADS]: The workflow involves recognized tools from established sources, such as k3s for local cluster testing and a vendor-managed Docker image (k8s-validator) for static manifest analysis.
- [COMMAND_EXECUTION]: The skill facilitates the use of standard Kubernetes CLI utilities including kubectl, helm, kustomize, and kapp. Security risk is mitigated by explicit instructions to always request user permission before running these commands, ensuring the agent does not perform unauthorized local system changes.
- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes user-provided Kubernetes artifacts. Ingestion points include manifest and Helm chart reading in references/manifests.md. No explicit boundary delimiters are specified for the untrusted content. The capability inventory includes kubectl and kapp for resource deployment. However, sanitization is performed via static validation in references/validate-snouty.md, and the risk is naturally limited by the tool's intended role in a dedicated testing pipeline.
Audit Metadata