sifter

Pass

Audited by Gen Agent Trust Hub on Apr 30, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the membrane CLI tool to perform operations such as logging in, connecting to services, and executing actions. These are standard operations for the platform's intended use case.
  • [EXTERNAL_DOWNLOADS]: The instructions direct the user to install the @membranehq/cli package via NPM. This is a legitimate utility provided by the vendor for interacting with their services.
  • [DATA_EXFILTRATION]: While the skill manages data from an external project management tool (Sifter), it does so through a controlled interface. The documentation explicitly discourages manual credential handling, reducing the risk of accidental exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 30, 2026, 08:11 PM
Security Audit — agent-trust-hub — sifter