The Agent Skills Directory

[Prompt Injection] (HIGH): The skill is vulnerable to Indirect Prompt Injection because it handles untrusted data and has high-privilege write capabilities.
Ingestion points: Processes untrusted body and media_url data in WHATSAPP_SEND_MESSAGE and WHATSAPP_SEND_MEDIA.
Boundary markers: Absent. There are no instructions to the agent to treat this data as untrusted.
Capability inventory: Includes high-impact tools for sending messages and media to external contacts.
Sanitization: No input validation or sanitization is mentioned.
[External Downloads] (MEDIUM): Relies on an unverified third-party MCP server (https://rube.app/mcp) which is not within the trusted organization scope.
[Data Exfiltration] (LOW): The tool WHATSAPP_SEND_MEDIA accepts a URL, which could be exploited for Server-Side Request Forgery (SSRF) to access internal network metadata if the agent environment is not isolated.

whatsapp-automation