The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted external data (messages and phone numbers) through its call actions. This represents an indirect prompt injection surface where a malicious input could potentially influence the agent's behavior during a call session. * Ingestion points: The message and to parameters in initiate_call, continue_call, and speak_to_user actions within SKILL.md. * Boundary markers: No specific delimiters or safety instructions are defined to separate user data from system instructions. * Capability inventory: The skill allows the agent to initiate and manage voice calls via external providers like Twilio or Plivo. * Sanitization: There is no documentation of sanitization or filtering of the message content before it is processed by the voice-call plugin.

voice-call