The Agent Skills Directory

Indirect Prompt Injection (LOW): The helper script scripts/generate_k8s_manifests.py contains a vulnerability surface where untrusted data could influence generated infrastructure manifests.
Ingestion points: CLI arguments such as --app-name and --image are interpolated into YAML templates.
Boundary markers: Absent.
Capability inventory: The script can write generated content to the local filesystem via the --output argument.
Sanitization: None; input is directly formatted into strings.
External Downloads (LOW): The documentation in examples/k8s-argocd/README.md instructs users to download and apply a manifest from an external source.
Source: https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml.
Execution: Applied via kubectl apply.
Status: argoproj is the official repository for ArgoCD but is not included in the predefined trusted organization list.
Safe Practices (SAFE): The validation script scripts/validate_deployment.py correctly uses yaml.safe_load_all to prevent YAML-based deserialization vulnerabilities.

deploying-applications