salesforce-marketing-cloud-mcp

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions for users to download components from external sources.
  • Fetches the MCP server source code from a GitHub repository (vinkius-labs/salesforce-marketing-mcp).
  • Provides instructions to pull a Docker image from Docker Hub (vinkius/salesforce-marketing-mcp).
  • Uses npx mcpfusion for deployment to a cloud environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection by processing external data.
  • Ingestion points: Data is retrieved from Salesforce Marketing Cloud via the query_data_extension and manage_subscriber tools.
  • Boundary markers: None are specified in the instructions to distinguish between data and instructions.
  • Capability inventory: The skill allows the agent to trigger marketing journeys (trigger_journey) and modify subscriber data (manage_subscriber), which could be abused if malicious instructions are found within a queried Data Extension.
  • Sanitization: No explicit sanitization or validation of the retrieved data is mentioned before processing.
  • [COMMAND_EXECUTION]: The documentation includes various shell commands for setup and troubleshooting.
  • Provides git clone, npm install, and docker run commands for local installation.
  • Includes a curl command for manual authentication testing against Salesforce APIs.
  • These commands are provided as instructions for the user and are not executed autonomously by the skill code.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 09:36 PM