openai-api

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The skill references the official openai Node.js package from a trusted organization.\n- Data Exposure & Exfiltration (SAFE): Code templates correctly utilize environment variables for authentication and target official OpenAI endpoints.\n- Indirect Prompt Injection (SAFE): The skill provides patterns for processing external data (text, images, audio). \n
  • Ingestion points: User messages in templates like chat-completion-basic.ts and planned vision/audio templates.\n
  • Boundary markers: Documentation explicitly mentions structured outputs and JSON schema validation.\n
  • Capability inventory: SDK calls to api.openai.com.\n
  • Sanitization: Recommends JSON schema validation for responses.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:07 PM