subconscious-dev

Pass

Audited by Gen Agent Trust Hub on Feb 20, 2026

Risk Level: SAFE
Full Analysis
  • [Data Exposure & Exfiltration] (SAFE): The documentation correctly identifies best practices for handling sensitive information. API keys are represented by placeholders (e.g., 'YOUR_API_KEY') or references to environment variables (e.g., 'process.env.SUBCONSCIOUS_API_KEY'), which prevents accidental exposure.
  • [Remote Code Execution] (SAFE): The skill describes a 'Function Tools' mechanism where the platform executes HTTP requests to user-defined URLs. While this involves outbound network activity, it is a documented core feature of the platform for tool integration and does not constitute an unauthorized RCE vulnerability within the skill itself.
  • [Prompt Injection] (SAFE): The provided files are reference documents and do not contain executable prompts or instructions intended to override AI safety guardrails or system instructions.
  • [Command Execution] (SAFE): Shell examples provided are standard curl commands for API interaction and do not include malicious payloads or unauthorized system access patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 20, 2026, 11:42 PM