devils-advocate

Pass

Audited by Gen Agent Trust Hub on Sep 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust safety orchestration layer that intercepts AI agent plans and actions to perform risk assessments before execution.
  • [SAFE]: The 'Gate Protocol' ensures the AI agent remains subordinate to the user, requiring an explicit '✅ Proceed' command for any operation involving side effects like file modifications, command execution, or deployments.
  • [SAFE]: The 'Building Protocol' mandates secure and high-quality coding practices, specifically requiring en_US identifiers and prohibiting hardcoded credentials or empty error handlers.
  • [SAFE]: The framework includes detailed sub-protocols like 'Handbrake' and 'Immediate Report' to escalate critical findings (e.g., PII exposure, SQL injection) to the user immediately during the analysis process.
  • [SAFE]: The skill contains explicit instructions ('Analyzed Content Boundary') to treat all analyzed content as untrusted input, preventing prompt injection attacks from modifying the skill's own security protocols.
  • [SAFE]: No evidence of data exfiltration, hidden network requests, or obfuscated code was found in the instructions or frameworks.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 22, 2026, 03:54 PM
Security Audit — agent-trust-hub — devils-advocate