zenscrape

Pass

Audited by Gen Agent Trust Hub on Apr 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the membrane CLI tool to execute actions and proxy requests to the Zenscrape API. This is expected behavior for an integration on the Membrane platform.\n- [EXTERNAL_DOWNLOADS]: The instructions direct the user to install the @membranehq/cli package from npm. This is a verified vendor resource for the Membrane ecosystem.\n- [PROMPT_INJECTION]: The skill processes data scraped from external websites, which introduces a risk of indirect prompt injection if the content contains malicious instructions for the agent.\n
  • Ingestion points: Web content retrieved via membrane request or membrane action run.\n
  • Boundary markers: None present.\n
  • Capability inventory: Shell command execution via the membrane CLI.\n
  • Sanitization: No data sanitization or filtering is specified in the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 2, 2026, 02:22 AM