The Agent Skills Directory

[SAFE]: The skill provides comprehensive instructions for the Better Auth library. It actively promotes security hardening by documenting the use of environment variables for secrets, implementing PKCE for OAuth flows, and enforcing CSRF protection.
[COMMAND_EXECUTION]: Mentions the use of the official @better-auth/cli tool for database migrations and schema generation. These operations are standard for the library's intended use case and are executed by the developer to maintain the authentication database.
[INDIRECT_PROMPT_INJECTION]: The skill describes handling external data from authentication requests. This surface is necessary for the library's functionality, and the documentation provides extensive guidance on security sanitization, including rate limiting, origin validation, and encrypted session tokens.
Ingestion points: Integration handlers in references/framework-integrations.md and hook contexts in references/hooks-and-middleware.md that process raw request headers and bodies.
Boundary markers: The documentation instructs developers to configure trustedOrigins and enables CSRF validation by default.
Capability inventory: The skill documents interactions with databases via ORM adapters and network operations for OAuth providers and communication services (email/SMS).
Sanitization: The library includes built-in protections against account enumeration (dummy timing operations) and uses JWE (JSON Web Encryption) for session data protection.

better-auth