indicator-design

Pass

Audited by Gen Agent Trust Hub on May 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No prompt injection or safety bypass techniques detected in the instructions.
  • [SAFE]: No sensitive data exposure or unauthorized network operations. The skill operates locally within the agent's context using standard library imports.
  • [SAFE]: No obfuscation, hidden commands, or malicious encoding found in the content.
  • [SAFE]: No remote code execution or unverifiable dependencies. It suggests the use of the well-known 'pandas' library for data processing.
  • [SAFE]: No persistence or privilege escalation mechanisms are present.
  • [SAFE]: Indirect prompt injection risk is minimal; the skill processes user-provided trading hypotheses to generate code templates but lacks access to high-risk tools or external data sources.
Audit Metadata
Risk Level
SAFE
Analyzed
May 10, 2026, 02:46 AM
Security Audit — agent-trust-hub — indicator-design