The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional Markdown content and code examples intended to guide a developer or AI agent in creating better CLI tools. All scripts provided are illustrative snippets within documentation rather than executable payloads.
[SECURITY_PRACTICES]: The skill actively promotes security by including rules like 'Never Read Secrets from Environment Variables' (rules/cli-secrets-not-envvars.md). This demonstrates a defensive posture by advising against common pitfalls like secret exposure in process lists (ps aux) or shell history.
[COMMAND_EXECUTION]: While the documentation contains shell command examples (e.g., curl, sudo mv, chmod), these are provided as standard installation templates for users to follow when distributing their own tools. There are no instructions for the agent to execute dangerous or unauthorized commands on the host system.
[REMOTE_CODE_EXECUTION]: There are no patterns of fetching and piping remote scripts into interpreters (e.g., curl|bash). The distribution section (rules/cli-distribution-binary.md) explicitly discusses downloading binaries, but does so within the context of documenting distribution best practices for the developer.

platform-cli