The Agent Skills Directory

[SAFE]: The skill enforces strict security controls for financial operations by requiring explicit user confirmation for all value-moving actions, including deposits, withdrawals, and credit top-ups.
[SAFE]: Sensitive management tokens are handled with appropriate care, with instructions mandating they be stored only in a secure secret store and never transmitted to unauthorized hosts or displayed in logs.
[SAFE]: The skill proactively identifies external data sources, such as agent session objectives and collaboration artifacts, as untrusted content. It includes specific instructions to prevent the agent from following links or executing instructions found within these sources, effectively mitigating risks associated with indirect prompt injection.

harness