ai-streaming

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill offers legitimate instructional content and code samples for building AI-powered features using industry-standard tools and patterns.
  • [COMMAND_EXECUTION]: Uses basic shell commands (cat, grep, find) in Phase 1 to analyze the project's current state and dependencies. These operations are benign and restricted to project discovery.
  • [EXTERNAL_DOWNLOADS]: Recommends the installation of official, well-known packages from the Vercel AI SDK ecosystem (ai, @ai-sdk/anthropic, @ai-sdk/openai) via the standard package manager.
  • [SAFE]: Provides security-conscious guidance for credential management by instructing users to add placeholders to a .env.example file rather than hardcoding actual keys, which is a recognized best practice.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 08:47 AM
Security Audit — agent-trust-hub — ai-streaming