The Agent Skills Directory

[SAFE]: The skill provides guidelines and local validation tools for software engineering. Analysis of all scripts and documentation confirms the absence of prompt injection, data exfiltration, or persistence mechanisms.- [COMMAND_EXECUTION]: Contains Python scripts in the scripts/ directory designed to scan local project files for architectural compliance and code coupling. These scripts operate locally, do not perform network requests, and use standard libraries for regex-based parsing.- [CREDENTIALS_UNSAFE]: The provided configuration template in assets/config-template.yaml follows security best practices by using shell-style placeholders like ${DB_PASSWORD} and instructing users to load secrets from environment variables rather than hardcoding them.

architect-first