The Agent Skills Directory

[SAFE]: The skill implements robust security controls by requiring regex validation (e.g., ^[a-z][a-z0-9-]*$) for user-supplied strings before they are used in shell commands like mkdir or cp, effectively preventing command injection.
[SAFE]: Instructions include an explicit safety directive to treat external content, such as Jira issue comments or existing files, as non-executable data rather than instructions, which mitigates indirect prompt injection risks.
[SAFE]: The use of shell commands is restricted to a specific allowlist (git clone, git pull, mkdir, cp) necessary for its stated purpose of repository management, following the principle of least privilege.

starting-breakdown