The Agent Skills Directory

[COMMAND_EXECUTION]: The Python scripts scripts/agent.py and scripts/process.py execute Helm CLI commands and security scanners using the subprocess module. These calls use list-based arguments and avoid shell invocation, preventing potential command injection vulnerabilities.
[EXTERNAL_DOWNLOADS]: The skill references the helm-secrets plugin for installation from GitHub. GitHub is recognized as a well-known service, and the plugin is a standard security tool in the Kubernetes ecosystem, making this reference safe under the trust scope rules.
[INDIRECT_PROMPT_INJECTION]: The skill exposes an attack surface by processing external Helm charts provided by the user. Ingestion points: Chart paths and files passed to the audit scripts. Boundary markers: No explicit markers are used in the scripts to delimit untrusted data for an LLM. Capability inventory: The scripts are capable of running system commands via the Helm CLI. Sanitization: The scripts perform regex-based auditing on raw YAML content without prior sanitization, which is consistent with the primary purpose of a security scanner.

securing-helm-chart-deployments