skills/samhvw8/dotfiles/pug-ui/Gen Agent Trust Hub

pug-ui

Warn

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill's templates download the Pug compiler from an unofficial GitHub Pages URL (https://gnjo.github.io/pug/pug.js) rather than an official project repository or a verified package registry.
  • Affected files: assets/template-bulma.html, assets/template-daisyui.html, assets/template-minimal.html, assets/template-pico.html.
  • [EXTERNAL_DOWNLOADS]: Fetches CSS framework resources from well-known services and CDNs including jsDelivr, Tailwind CSS, and Simple.css.
  • [COMMAND_EXECUTION]: The templates use client-side compilation via pug.render() and inject the resulting HTML into the DOM using innerHTML. This allows for the execution of any scripts or event handlers that may be present in the generated Pug source.
  • [PROMPT_INJECTION]: The skill establishes a surface for indirect prompt injection by rendering content generated by the AI based on user prompts.
  • Ingestion points: The pug-src script block in all generated HTML templates serves as the entry point for untrusted data.
  • Boundary markers: The template lacks markers or instructions to prevent the compiler from processing malicious scripts embedded within the generated Pug code.
  • Capability inventory: The skill has the capability to execute browser-side scripts via the innerHTML assignment of compiled Pug output.
  • Sanitization: No sanitization, escaping, or validation is performed on the Pug source before it is rendered into the browser context.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 29, 2026, 02:10 PM
Security Audit — agent-trust-hub — pug-ui