The Agent Skills Directory

[SAFE]: The skill implements a legitimate iterative development workflow using standard shell commands and git operations. All operations are local to the project directory, and no unauthorized network activity or data exfiltration patterns were observed.- [SAFE]: Prompt templates are structured to provide specific task instructions and maintain focus on the implementation spec. The inclusion of explicit rules against modifying project meta-configuration or pushing to remote repositories enhances the safety of the skill execution.- [SAFE]: The skill relies on standard project files (e.g., package.json, AGENTS.md) and does not introduce unverifiable external dependencies or execute remote scripts.

ralph