workflow-state-machine-simulation

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a workflow for generating visualization artifacts (Mermaid state diagrams and interactive HTML files) without executing the target source code, ensuring a clear separation between modeling and execution.
  • [SAFE]: The interactive HTML component is designed to be self-contained, using native browser APIs and SVG for rendering. It explicitly avoids external dependencies, network requests, and dangerous functions like eval() or Function() constructors.
  • [SAFE]: Data handling instructions follow security best practices by treating target-derived strings as untrusted data and requiring safe injection into the DOM using textContent or equivalent APIs to prevent cross-site scripting (XSS).
  • [SAFE]: Potential indirect prompt injection risks from processing external target files are mitigated by clear instructions to model the source authoritatively without executing it and to sanitize all displayed content.
  • [SAFE]: The generation process enforces template invariance for the HTML artifact, ensuring that only declarative machine data is updated while the underlying visualization logic remains static and verified.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 01:01 PM
Security Audit — agent-trust-hub — workflow-state-machine-simulation