The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources, including CRM leads and email correspondence (Step 1). This creates a surface for indirect prompt injection if an attacker-controlled lead sends an email containing malicious instructions. However, the risk is low due to explicit 'approval gates' (Points of Confirmation) that prevent the agent from taking autonomous actions like sending emails, booking calendars, or changing CRM states without human review. The skill also lacks capabilities for code execution or arbitrary file writes.
[SAFE]: The skill demonstrates security best practices by implementing manual confirmation steps for all outbound actions (emails and calendar bookings). All data processing is localized to the user's business context.

call-list