The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run development commands such as bundle install, rails generate, and bundle exec rspec. These are standard tools for Ruby on Rails development and are used within the scope of the skill's purpose for implementing and testing authorization frameworks.
[EXTERNAL_DOWNLOADS]: The skill references the installation of the pundit and cancancan gems via Bundler. These are well-known, industry-standard libraries for authorization in the Ruby ecosystem.
[INDIRECT_PROMPT_INJECTION]: The skill involves the agent reading and analyzing existing project code to implement authorization logic.
Ingestion points: Project source files are inspected to determine existing authorization frameworks and conventions.
Boundary markers: None explicitly defined in the instructions for code analysis isolation.
Capability inventory: File system writes (creating policy and ability files), command execution (rails generate, bundle install), and local network testing (curl).
Sanitization: The skill mandates rigorous verification through automated RSpec tests and manual console output of denied actions to ensure the integrity of the implemented logic.

implement-authorization