The Agent Skills Directory

[SAFE]: The skill provides educational documentation, configuration templates, and example commands for zero-trust security implementations. The content is strictly technical and follows industry best practices for modern security architecture.
[COMMAND_EXECUTION]: The skill includes example curl commands for interacting with the Cloudflare API. These commands appropriately use environment variables (e.g., ${CF_TOKEN}, ${ACCOUNT_ID}) to handle sensitive credentials, which is the standard method for avoiding hardcoded secrets.
[REMOTE_CODE_EXECUTION]: The Kubernetes deployment templates reference a standard container image (quay.io/oauth2-proxy/oauth2-proxy:v7.6.0) from a well-known public registry. This is an expected and routine practice for infrastructure-as-code and DevOps skills.
[INDIRECT_PROMPT_INJECTION]: The skill functions as a repository for templates and architectural guidance. It does not ingest or process untrusted external data during its execution context, minimizing the risk of indirect injection.

zero-trust