The Agent Skills Directory

[SAFE]: The skill provides comprehensive guidance for Rails authentication that prioritizes security best practices such as utilizing ActiveSupport::SecurityUtils.secure_compare to mitigate timing attacks during token verification.
[SAFE]: No evidence of prompt injection or malicious instructions was found; the instructions are consistently focused on the stated goal of implementing a passwordless authentication flow.
[SAFE]: Data protection is addressed through the recommended use of filter_parameter_logging.rb to ensure sensitive tokens and codes are not leaked into application logs.
[SAFE]: The skill uses cryptographic message verifiers to bind authentication codes to specific email addresses, preventing attackers from using a valid code to sign in as a different user.
[SAFE]: The configuration of allowed-tools is restricted to standard Rails development commands (generate, db, test), providing a narrow and safe interface for the AI agent to interact with the codebase.
[SAFE]: The skill includes defensive measures against accidental data leakage, such as an after_action guard that prevents flashing authentication codes in non-development environments.

magic-link-auth