The Agent Skills Directory

[SAFE]: The skill incorporates a 'security-reviewer' role specifically tasked with tracing untrusted inputs to enforcement boundaries and inspecting logs for credential leaks.
[SAFE]: It implements strict data handling rules, explicitly instructing that private tokens should not be printed during CLI operations and that credentials must be kept out of project artifacts.
[SAFE]: The framework reinforces the principle of least privilege by stating that role assignment does not expand the agent's authorization and requires explicit user approval for risky actions like store submissions or paid research.

mobile-app-agency