The Agent Skills Directory

Persistence Mechanism via Shell Profiles: The instructions suggest adding environment variables to shell configuration files such as ~/.bashrc and ~/.zshrc. While this is a common method for maintaining configuration across terminal sessions, modifying these files is a form of persistence that should be reviewed to ensure only intended variables are added.
Indirect Prompt Injection Attack Surface: The skill enables the agent to fetch and process data from the external Google Ads API (e.g., campaign names and reporting metrics). Since this data originates from an external source that could potentially contain untrusted content, there is a possibility for indirect prompt injection if the agent interprets embedded instructions within that data.
Credential Handling Considerations: The setup process requires the user to input and store sensitive OAuth 2.0 credentials, including Client Secrets and Refresh Tokens. The skill correctly recommends using environment variables rather than hardcoding them in scripts, which aligns with security best practices for credential management.

google-ads-api-mcp-setup