design-an-interface

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists entirely of instructional markdown and configuration metadata. It contains no executable scripts, binary components, or remote code dependencies.
  • [NO_CODE]: No custom code or external package dependencies (Python/Node.js) are utilized. The functionality relies on the agent's native reasoning and standard orchestration tools.
  • [PROMPT_INJECTION]: The skill uses a template-based approach to pass user-provided module descriptions and requirements to sub-agents. While this represents a surface for indirect prompt injection, it is the intended functionality of the generative design process and lacks the capabilities (such as network or file system access) to be exploited maliciously.
  • [COMMAND_EXECUTION]: No shell commands, system calls, or privilege escalation attempts (e.g., sudo) were found in the instructions or metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 10:25 PM
Security Audit — agent-trust-hub — design-an-interface