The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use standard directory discovery tools like tree and ls. These operations are limited to local, non-recursive exploration for the purpose of documenting directory contents.
[COMMAND_EXECUTION]: Specific read-only git commands (git status, git ls-files, git log) are used to gather repository metadata. The instructions explicitly use pipes to cat to prevent interactive shell sessions and forbid any state-changing operations like commit or push.
[DATA_EXFILTRATION]: While the skill reads local documentation (e.g., root and parent READMEs) to gain context, it does not include any network-capable tools or patterns for transmitting this data to external servers.
[SAFE]: The skill does not employ obfuscation, hardcoded credentials, remote code execution, or privilege escalation. All external references point to internal assets within the skill's own directory structure.

readme-creation