box
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill interacts with the Box API via a local binary located at
/opt/hatch/bin/box-cli. This tool is used to perform searches, read files, and manage account status.\n- [EXTERNAL_DOWNLOADS]: The skill performs legitimate network operations to official Box domains (api.box.comandmcp.box.com) to facilitate file management and synchronization.\n- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it allows the agent to ingest data from an external source (Box files).\n - Ingestion points: File content is retrieved using the
get_file_contenttool as described inSKILL.md.\n - Boundary markers: No specific boundary markers or instructions to ignore embedded commands are present in the skill's instructions.\n
- Capability inventory: The agent has permissions to perform impactful actions such as
upload_file,delete_file, andsharing.manage.\n - Sanitization: The skill does not describe any specific sanitization procedures for data read from the Box service.
Audit Metadata