interactive-study-guide

Warn

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [Command Execution] (MEDIUM): The skill executes a local shell script (scripts/scaffold.sh) and starts a development server via npx vite. These commands use project-specific arguments that could be subject to injection if the input data or project metadata is maliciously crafted.
  • [External Downloads] (MEDIUM): The workflow involves running npm install and npx vite, which download and execute code from the public npm registry. This introduces risks related to supply chain attacks and untrusted dependency execution.
  • [Indirect Prompt Injection] (LOW): The skill processes untrusted markdown files as its primary input. 1. Ingestion points: User-provided markdown file path in Step 1. 2. Boundary markers: None; parsing depends on expected heading patterns. 3. Capability inventory: Local bash script execution, npm package installation, and Vite server execution. 4. Sanitization: No sanitization or validation of the input markdown content is described before it is transformed into application code.
  • [Dynamic Execution] (MEDIUM): The skill is designed for dynamic code generation, creating a Vite-based single-page application and individual interactive components based on the input data, then executing this generated code via a local development server.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 21, 2026, 02:06 PM