intercom
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data from Intercom via Apideck's Proxy API, which represents a potential attack surface for indirect prompt injection.\n
- Ingestion points: Raw HTTP responses from the Intercom API fetched through https://unify.apideck.com/proxy (SKILL.md).\n
- Boundary markers: No specific boundary markers or instruction-delimiting tags are mentioned for processing external responses.\n
- Capability inventory: The skill enables network communication with the Intercom API, allowing for data retrieval and modification (SKILL.md).\n
- Sanitization: There is no mention of explicit sanitization or filtering of the downstream API responses before they are presented to the agent.\n- [EXTERNAL_DOWNLOADS]: The skill directs network traffic to https://unify.apideck.com/proxy. This is the official endpoint for the skill vendor's proxy service and is used as intended to securely handle API requests to Intercom without exposing credentials.\n- [COMMAND_EXECUTION]: The documentation includes a curl command example to demonstrate how the agent should structure requests to the Proxy API. This is provided for informational purposes to guide the agent's interaction with the service.
Audit Metadata