nano-nets

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes data from Nano Nets, which could include untrusted content leading to indirect prompt injection.
  • Ingestion points: Retrieves notes, files, and activity logs from the Nano Nets platform (SKILL.md).
  • Boundary markers: No specific delimiters or instructions are used to distinguish between data and instructions during prompt construction.
  • Capability inventory: The skill can execute predefined actions and make arbitrary API requests through a proxy (SKILL.md).
  • Sanitization: There is no evidence of data sanitization or validation of the content retrieved from the API.
  • [COMMAND_EXECUTION]: Core functionality is implemented through shell commands.
  • Evidence: Uses the 'membrane' CLI for login, action execution, and API requests.
  • [EXTERNAL_DOWNLOADS]: Requires the installation of a global CLI package.
  • Evidence: Instructions include 'npm install -g @membranehq/cli'. This is an official vendor package used to facilitate the integration.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 06:30 AM