The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs users to add an external MCP server at https://rube.app/mcp. This domain is not identified as a trusted organization or well-known service in the provided lists, representing an unverified remote dependency.- [DATA_EXFILTRATION]: All Postmark operations, including sending batch emails and retrieving bounces, are routed through the rube.app infrastructure. This creates a significant risk of data exposure where sensitive information such as recipient email addresses and message content are processed by an external third-party proxy.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface. Ingestion points: External data enters the context via POSTMARK_GET_TEMPLATE and POSTMARK_LIST_TEMPLATES. Boundary markers: There are no defined delimiters or instructions to ignore embedded commands within templates. Capability inventory: The skill allows for sending batch emails (POSTMARK_SEND_BATCH_WITH_TEMPLATES) and modifying server configurations (POSTMARK_EDIT_SERVER). Sanitization: No evidence of content escaping or validation is present to prevent malicious instructions within templates from influencing the agent's behavior.

postmark-automation