The Agent Skills Directory

[SAFE]: The skill performs standard local project management tasks including reading project files and git history to establish context. It writes validated designs to the docs/plans/ directory and manages project status within a .planning/ directory. While reading project files introduces a surface for indirect prompt injection, this behavior is essential for the skill's purpose and occurs within a trusted local context.
[SAFE]: Command execution is restricted to standard git operations and a specific local initialization script (init-planning-dir.sh) located within the plugin's root directory. These actions are used to maintain project structure and do not involve untrusted remote sources or requests for elevated privileges.
[SAFE]: The skill operates entirely within the local environment and does not perform any network communication, ensuring that project data remains private and no data exfiltration occurs.

brainstorming