unbox

Warn

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing the @membranehq/cli package globally using npm. This is a standard requirement for the Membrane platform.
  • [COMMAND_EXECUTION]: Executes several membrane CLI commands for authentication, connection establishment, and action execution.
  • [METADATA_POISONING]: The skill includes a documentation link to Apple's ARKit documentation (developer.apple.com/documentation/arkit), which is unrelated to the Unbox subscription platform described. This constitutes misleading metadata.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by processing external data. 1. Ingestion points: Data retrieved from Unbox connections via CLI commands. 2. Boundary markers: Not present in the instructions. 3. Capability inventory: The skill can execute commands and run actions through the Membrane CLI. 4. Sanitization: No sanitization or validation of external input is described.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 28, 2026, 11:51 PM