phrase

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security risks or malicious patterns were identified in the skill instructions or dependencies.\n- [EXTERNAL_DOWNLOADS]: The skill references the installation of the official @membranehq/cli package from npm, which is a trusted vendor resource.\n- [COMMAND_EXECUTION]: The skill utilizes the membrane CLI to interact with Phrase, performing authenticated actions and data retrieval as intended.\n- [PROMPT_INJECTION]: The skill processes data from the Phrase API, creating an indirect prompt injection surface. Ingestion points: Phrase API responses (SKILL.md). Boundary markers: Not specified. Capability inventory: membrane action run (e.g., create-project, update-locale). Sanitization: Relies on Phrase API and Membrane platform security layers.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 11:05 PM