The Agent Skills Directory

[SAFE]: The skill implements a highly defensive Git workflow. It specifically instructs the AI agent to avoid committing sensitive files like .env or credentials.json, minimizing the risk of accidental data exposure.
[COMMAND_EXECUTION]: Git commands are used for their intended purpose. The instructions emphasize safety by forbidding destructive flags like --no-verify and ensuring that hooks are run. It also uses sandbox-compatible methods for multi-line commit messages using printf and $TMPDIR to avoid potential environment-specific execution failures.
[SAFE]: No remote code execution, obfuscation, or persistence mechanisms were found. All scripts provided are local examples and do not fetch external executable content.

commit-push-sync