alphasense

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the @membranehq/cli package from the NPM registry to interact with the platform infrastructure.- [COMMAND_EXECUTION]: Utilizes various shell commands including membrane login, membrane connect, and membrane action run to manage the integration and execute market intelligence tasks.- [PROMPT_INJECTION]: The skill processes data from AlphaSense including company filings and news which represents a potential surface for indirect prompt injection though no active exploitation was detected.- [SAFE]: Delegating authentication to the Membrane platform is a positive security practice that prevents the need for manual API key handling or credential exposure within the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 02:48 AM
Security Audit — agent-trust-hub — alphasense