The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/scan_repo.py executes git via subprocess.run to determine which files have changed relative to the current HEAD. This allows the tool to scope its analysis to the current diff. The command is implemented using a list of arguments rather than a shell string, which is a secure practice that prevents command injection.
[SAFE]: The skill performs local static analysis of project files (source code and manifests like package.json) to identify technology stacks and architectural violations. These operations are performed using the Python standard library without any network calls or external data exfiltration.
[SAFE]: The skill provides integration templates in the assets/ directory, including Git pre-commit hooks and IDE settings. These are benign automation tools that facilitate context loading and local checks, requiring manual setup by the user.

clean-code