The Agent Skills Directory

[SAFE]: The skill content is instructional and focused on security best practices, such as preventing credential exposure and blocking dangerous shell commands via hooks.
[COMMAND_EXECUTION]: Includes examples of standard shell commands for project initialization (/init), file creation (cat), and directory management (mkdir), which are appropriate for its stated purpose of environment configuration.
[EXTERNAL_DOWNLOADS]: References the use of npx to run an auditing tool (cc-safe) and git for sharing configurations, which are standard development workflows.
[PROMPT_INJECTION]: Proactively addresses prompt injection by providing templates for security hooks designed to detect and block malicious command patterns like rm -rf / or curl | sh.

agent-configuration