interactive-study-guide

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands to scaffold a project, install dependencies, and start a development server. It specifically instructs the agent to run bash scripts/scaffold.sh, npm install, and npx vite --open within the project directory.\n- [EXTERNAL_DOWNLOADS]: The skill fetches resources from well-known services. It installs d3 and prismjs from the npm registry and loads fonts from Google's CDN (fonts.googleapis.com). \n- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it parses user-provided markdown files to generate application code. \n
  • Ingestion points: Step 1 in SKILL.md (Locate and Parse the Study Guide).\n
  • Boundary markers: Absent; uses regex and heading patterns for parsing without explicit delimiters for embedded content.\n
  • Capability inventory: Shell execution (bash, npm, npx) and file system writes.\n
  • Sanitization: Absent; the skill transforms markdown sections into HTML/JS components based on the content provided.\n- [DYNAMIC_EXECUTION]: The skill dynamically generates a web application, including HTML, CSS, and JavaScript, at runtime. It uses templates in references/section-strategies.md and design tokens in references/design-system.md to build the interactive experience.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 01:25 PM
Security Audit — agent-trust-hub — interactive-study-guide