prompt-hub

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill incorporates proactive security measures as 'load-bearing invariants,' specifically requiring a 'Red Team' status check and an automated 'Voice gate' (lib/voice/frankx-voice.ts) to audit outputs for banned content.
  • [SAFE]: External references are limited to well-known technology providers and official documentation from trusted organizations such as Anthropic, Google, and OpenAI, as well as established open-source tools like Promptfoo and Fabric.
  • [SAFE]: The skill uses a structured routing mechanism via a conductor agent, reducing the risk of unauthorized tool use or accidental exposure by enforcing canonical workflows.
  • [SAFE]: No instances of prompt injection, obfuscation, or data exfiltration attempts were found during the analysis of the skill's logic and instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 06:40 AM
Security Audit — agent-trust-hub — prompt-hub