The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses dynamic context injection in SKILL.md to execute git status, git diff HEAD, and git log when the skill is loaded. These are read-only operations used to provide the agent with repository context and do not involve network access or the injection of user-supplied arguments into the shell.
[COMMAND_EXECUTION]: The skill's primary workflow involves executing git add and git commit commands. The instructions specifically advise against using global flags like git add . to prevent the accidental staging of unwanted or sensitive files.
[SAFE]: The skill includes explicit security rules, such as NEVER commit secrets and instructions to abort the process if staged files are found to contain credentials or sensitive information.
[SAFE]: The skill logic is localized to git operations within the repository and does not attempt to access external networks, modify system configurations, or achieve persistence.

commit