The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs standard Rails development commands including bundle exec rspec for testing, bin/rails generate for scaffolding, and bin/rails console for manual verification. These tools are used appropriately within the primary scope of authorization management and do not exhibit malicious patterns.
[SAFE]: The skill implements a defensive security posture by explicitly instructing the agent to 'deny by default' and verify authorization in every controller action. It includes a built-in audit command using grep to identify potential security gaps in controllers without calling external services.

policy-agent