The Agent Skills Directory

[SAFE]: Naming Conventions and Best Practices: The skill provides structural guidance for naming Invoke tasks and organizing Python automation modules. These instructions are purely documentation-oriented and do not trigger any execution risk.
[SAFE]: Security-Minded Configuration Advice: The instructions include a specific section on preventing sensitive data exposure. It warns developers about a known behavior in pyinvoke where c.run can echo stdin (including passwords) to the terminal output. It correctly recommends using sudo -n and non-interactive environment variables (DEBIAN_FRONTEND=noninteractive) as mitigations.
[SAFE]: Environment Configuration Context: The skill mentions an 'allowlist' for Claude Code that auto-approves certain read-only command patterns (e.g., status, check, list). This is documented as an existing workflow configuration in the author's own repositories (power-user-linux-setup) and does not represent an attempt to override the current agent's safety protocols or permissions.
[SAFE]: Tool Wiring and Maintenance: The 'Wiring traps' section provides technical advice on how Invoke's Collection.from_module behaves, helping developers avoid unintended task exposure through Python imports. This is standard development practice for the Invoke library.

invoke-task-conventions-taudelta