The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the implementation to load the official Google Identity Services SDK from https://accounts.google.com/gsi/client. This is a standard, well-known, and safe resource for implementing Google authentication.
[REMOTE_CODE_EXECUTION]: The backend instructions recommend using official, well-known libraries such as google-auth for Python and google-auth-library for Node.js. These are standard packages from a trusted provider for the intended purpose.
[CREDENTIALS_UNSAFE]: The skill follows security best practices for credential handling. It correctly identifies the GOOGLE_CLIENT_ID as a public identifier and explicitly warns against including the client secret in the application. It also provides excellent guidance on using HttpOnly, Secure, and SameSite=Strict flags for session cookies.
[DATA_EXFILTRATION]: No malicious data exfiltration patterns were detected. The network operations described are limited to official Google verification endpoints and standard application backend communication.
[PROMPT_INJECTION]: The skill instructions do not contain any patterns intended to override agent behavior, bypass safety filters, or extract system prompts.

google-signin