enforce-policy-as-code
Warn
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches the gator and kyverno-cli tools from the official release pages of the Open Policy Agent and Kyverno GitHub organizations.- [COMMAND_EXECUTION]: Employs sudo to move and install binaries into /usr/local/bin/, requiring elevated host-level access during the setup process.- [COMMAND_EXECUTION]: Executes administrative kubectl and helm commands to deploy cluster-wide components, including admission controllers, webhooks, and custom resource definitions.- [REMOTE_CODE_EXECUTION]: Downloads and subsequently executes external binaries within CI/CD workflows and git hooks to validate Kubernetes manifests.- [COMMAND_EXECUTION]: Vulnerable to indirect prompt injection when processing Kubernetes manifests.- Ingestion points: The skill reads and evaluates Kubernetes manifest files (e.g., manifests/*.yaml) within CI/CD validation steps and pre-commit hooks.- Boundary markers: Absent; there are no specific delimiters or instructional guards to prevent the agent from obeying instructions embedded in the manifests.- Capability inventory: The skill utilizes bash, kubectl, sudo, and helm to interact with the cluster and local filesystem.- Sanitization: Absent; the content of the external manifests is processed directly by the CLI tools without prior validation or escaping.
Audit Metadata