md-slides

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references Google Fonts (fonts.googleapis.com) and optionally Prism.js from the jsDelivr CDN (cdn.jsdelivr.net) to provide typography and syntax highlighting in the generated presentations. These are well-known technology services and are used according to their intended public purpose.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied markdown content to generate HTML artifacts, which introduces a potential surface for indirect injection if the content is viewed in a browser.
  • Ingestion points: Markdown text is read from files or standard input by slide_splitter.py.
  • Boundary markers: The skill uses standard markdown separators (--- and # ) to define slides but does not implement specific delimiters to isolate user content from the rendering logic.
  • Capability inventory: The deck_html_renderer.py script possesses the capability to write HTML and JavaScript files to the local filesystem.
  • Sanitization: The renderer consistently applies html.escape() to titles, speaker notes, and code blocks before interpolation, mitigating cross-site scripting (XSS) risks in the generated output.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 07:19 AM
Security Audit — agent-trust-hub — md-slides