The Agent Skills Directory

[SAFE]: The skill content is purely instructional and provides templates for architectural workflows. It does not contain any executable scripts, obfuscated code, or network-based exfiltration patterns.
[COMMAND_EXECUTION]: The skill documentation suggests using the agent's built-in run_command tool for legitimate inspection tasks, such as auditing dependency trees with npm ls and pip show.
[DATA_EXPOSURE]: Includes instructions for the agent to use read_file and grep to understand existing project architectures, which is a standard requirement for its role as an architecture designer.
[NO_CODE]: No external scripts or code files are bundled with this skill; it relies entirely on the agent's pre-configured toolset.

architecture-designer