intercom-automation

Warn

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill directs users to add an external MCP server endpoint (https://rube.app/mcp) to their agent configuration. This server provides the underlying logic and execution environment for all Intercom tools used by the skill.
  • [DATA_EXFILTRATION]: By using the Rube MCP service, the skill transmits sensitive Intercom workspace data—such as conversation history, email addresses, and admin identifiers—to the third-party domain rube.app. The setup process includes an OAuth authentication flow handled by this external service.
  • [PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection because it retrieves and processes untrusted data from Intercom conversations.
  • Ingestion points: The tools INTERCOM_LIST_CONVERSATIONS and INTERCOM_GET_CONVERSATION ingest message content provided by external users and leads.
  • Boundary markers: There are no instructions or delimiters specified to prevent the agent from interpreting instructions embedded within retrieved Intercom messages.
  • Capability inventory: The agent possesses capabilities to perform state-changing actions based on processed data, such as INTERCOM_REPLY_TO_CONVERSATION, INTERCOM_ASSIGN_CONVERSATION, and INTERCOM_CLOSE_CONVERSATION.
  • Sanitization: The skill does not include steps for sanitizing, filtering, or validating message content before the agent processes it.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 11, 2026, 07:51 PM
Security Audit — agent-trust-hub — intercom-automation